report.object
AI agents for financial-institution service and operations
Expert-reviewed from public records, platform data, primary sources, and private research beyond AI search. Use /api/v1/reports/rpt_financial_services_member_service_ai_agents_2026_05/sample before purchase, then download PDF, JSON, or Markdown after checkout.
page.markdown
AI agents for financial-institution service and operations.md
# AI agents for financial-institution service and operations
Financial-institution service agents should be approved by obligation and authority: identity, payments, complaints, disputes, KYC, fraud, credit, insurance and system-of-record actions each need audit proof.
## Object
| field | value |
| ---------------- | ------------------------------------------------------- |
| report_id | rpt_financial_services_member_service_ai_agents_2026_05 |
| schema_version | 2026-05-28 |
| last_updated_at | 2026-06-17 |
| category | AGENTS |
| sector | Financial services |
| region | US/EU/UK |
| confidence_score | 0.87 |
| source_count | 73 |
| evidence_count | 77 |
## Endpoints
| action | endpoint |
| ------------ | ---------------------------------------------------------------------------------- |
| summary | /api/v1/reports/rpt_financial_services_member_service_ai_agents_2026_05 |
| sample | /api/v1/reports/rpt_financial_services_member_service_ai_agents_2026_05/sample |
| schema | /api/v1/reports/rpt_financial_services_member_service_ai_agents_2026_05/schema |
| quote | /api/v1/reports/rpt_financial_services_member_service_ai_agents_2026_05/quote |
| purchase | /checkout/rpt_financial_services_member_service_ai_agents_2026_05?view=agent |
| versions | /api/v1/reports/rpt_financial_services_member_service_ai_agents_2026_05/versions |
| changelog | /api/v1/reports/rpt_financial_services_member_service_ai_agents_2026_05/changelog |
| monitor | /api/v1/monitors?report_id=rpt_financial_services_member_service_ai_agents_2026_05 |
| mcp_resource | colophon://reports/rpt_financial_services_member_service_ai_agents_2026_05/sample |
## Coverage
| field | value |
| -------------- | ------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------ |
| category | AGENTS |
| sector | Financial services |
| region | US/EU/UK |
| topics | banking AI agents, financial services member service, credit union AI, consumer finance chatbots, bank contact centre AI, digital banking assistants, voice agents, complaint handling, Regulation E disputes, collections AI, GLBA, bank information security, credit union safeguards, DORA, ICT third-party risk, EBA ICT risk, third-party risk, core banking integrations, approved answers, AI audit logs, financial institution AI, KYC AI, AML AI, fraud investigation AI, broker-dealer generative AI, insurance AI governance, BaaS sponsor bank oversight, adverse action AI, Regulation S-P, remittance transfer AI, mortgage servicing AI, Consumer Duty AI |
| source_classes | regulator, regulation, standard, vendor_docs, case_study, privacy_terms, vendor_press, api_docs, pricing_page |
| update_cadence | Weekly |
## Sample JSON
```json
{
"report_id": "rpt_financial_services_member_service_ai_agents_2026_05",
"slug": "ai-agents-financial-institution-service-operations",
"schema_version": "2026-05-28",
"title": "AI agents for financial-institution service and operations",
"abstract": "Financial-institution service agents should be approved by obligation and authority: identity, payments, complaints, disputes, KYC, fraud, credit, insurance and system-of-record actions each need audit proof.",
"category": "AGENTS",
"sector": "Financial services",
"region": "US/EU/UK",
"topics": [
"banking AI agents",
"financial services member service",
"credit union AI",
"consumer finance chatbots",
"bank contact centre AI",
"digital banking assistants",
"voice agents",
"complaint handling",
"Regulation E disputes",
"collections AI",
"GLBA",
"bank information security",
"credit union safeguards",
"DORA",
"ICT third-party risk",
"EBA ICT risk",
"third-party risk",
"core banking integrations",
"approved answers",
"AI audit logs",
"financial institution AI",
"KYC AI",
"AML AI",
"fraud investigation AI",
"broker-dealer generative AI",
"insurance AI governance",
"BaaS sponsor bank oversight",
"adverse action AI",
"Regulation S-P",
"remittance transfer AI",
"mortgage servicing AI",
"Consumer Duty AI"
],
"last_updated_at": "2026-06-17T00:00:00Z",
"source_count": 73,
"evidence_count": 77,
"confidence_score": 0.87,
"web_url": "/r/ai-agents-financial-institution-service-operations/rpt_financial_services_member_service_ai_agents_2026_05",
"sample_endpoint": "/api/v1/reports/rpt_financial_services_member_service_ai_agents_2026_05/sample",
"inventory_endpoint": "/api/v1/reports/rpt_financial_services_member_service_ai_agents_2026_05/inventory",
"full_endpoint": "/api/v1/reports/rpt_financial_services_member_service_ai_agents_2026_05/full",
"schema_endpoint": "/api/v1/reports/rpt_financial_services_member_service_ai_agents_2026_05/schema",
"redaction": {
"status": "sample",
"paid_sections_redacted": true,
"preview_limits": [
"summary",
"table_of_contents",
"evidence_profile"
]
},
"coverage": {
"category": "AGENTS",
"sector": "Financial services",
"region": "US/EU/UK",
"topics": [
"banking AI agents",
"financial services member service",
"credit union AI",
"consumer finance chatbots",
"bank contact centre AI",
"digital banking assistants",
"voice agents",
"complaint handling",
"Regulation E disputes",
"collections AI",
"GLBA",
"bank information security",
"credit union safeguards",
"DORA",
"ICT third-party risk",
"EBA ICT risk",
"third-party risk",
"core banking integrations",
"approved answers",
"AI audit logs",
"financial institution AI",
"KYC AI",
"AML AI",
"fraud investigation AI",
"broker-dealer generative AI",
"insurance AI governance",
"BaaS sponsor bank oversight",
"adverse action AI",
"Regulation S-P",
"remittance transfer AI",
"mortgage servicing AI",
"Consumer Duty AI"
],
"source_classes": [
"regulator",
"regulation",
"standard",
"vendor_docs",
"case_study",
"privacy_terms",
"vendor_press",
"api_docs",
"pricing_page"
]
},
"confidence": {
"overall_confidence": 0.87,
"source_quality": 0.88,
"source_recency": 0.94,
"source_density": 0.9,
"claim_verifiability": 0.84,
"geographic_coverage": 0.76,
"known_gaps": [
"Independent end-to-end benchmarks for financial-institution service agents remain thin.",
"Vendor case studies report useful workflow and outcome signals but rarely expose false resolution, repeat contact, complaint handling, authentication failure, financial-crime boundary, servicing-error classification or reviewer-minute data.",
"Most contract terms, SOC reports, subprocessor schedules, model-provider routes, data retention settings and pricing details require buyer diligence under NDA.",
"US bank, credit-union, consumer-finance, BSA/AML, securities, insurance and UK conduct sources plus EU DORA and ICT/security anchors are included; local conduct, privacy, insurance, securities and outsourcing rules still need buyer-specific review."
]
},
"document": {
"object": "report_document",
"schemaVersion": "2026-05-28",
"title": "AI agents for financial-institution service and operations",
"subtitle": "The agent becomes institution-grade only when the route names the customer, product, obligation, system of record and owner.",
"abstract": "Financial-institution service agents should be approved by obligation and authority: identity, payments, complaints, disputes, KYC, fraud, credit, insurance and system-of-record actions each need audit proof.",
"category": "AGENTS",
"sector": "Financial services",
"region": "US/EU/UK",
"topics": [
"banking AI agents",
"financial services member service",
"credit union AI",
"consumer finance chatbots",
"bank contact centre AI",
"digital banking assistants",
"voice agents",
"complaint handling",
"Regulation E disputes",
"collections AI",
"GLBA",
"bank information security",
"credit union safeguards",
"DORA",
"ICT third-party risk",
"EBA ICT risk",
"third-party risk",
"core banking integrations",
"approved answers",
"AI audit logs",
"financial institution AI",
"KYC AI",
"AML AI",
"fraud investigation AI",
"broker-dealer generative AI",
"insurance AI governance",
"BaaS sponsor bank oversight",
"adverse action AI",
"Regulation S-P",
"remittance transfer AI",
"mortgage servicing AI",
"Consumer Duty AI"
],
"sources": [
{
"id": "S1",
"label": "CFPB - Chatbots in consumer finance",
"sourceClass": "regulator",
"accessedAt": "2026-06-01",
"status": "verified",
"access": "sample",
"transformChain": [],
"url": "https://www.consumerfinance.gov/data-research/research-reports/chatbots-in-consumer-finance/chatbots-in-consumer-finance/"
},
{
"id": "S2",
"label": "CFPB - The CFPB has entered the chat",
"sourceClass": "regulator",
"accessedAt": "2026-06-01",
"status": "verified",
"access": "sample",
"transformChain": [],
"url": "https://www.consumerfinance.gov/about-us/blog/cfpb-has-entered-the-chat/"
},
{
"id": "S3",
"label": "CFPB - UDAAP examination procedures",
"sourceClass": "regulator",
"accessedAt": "2026-06-01",
"status": "verified",
"access": "sample",
"transformChain": [],
"url": "https://files-prod.consumerfinance.gov/f/documents/cfpb_unfair-deceptive-abusive-acts-practices-udaaps_procedures_2023-09.pdf"
},
{
"id": "S4",
"label": "CFPB - Regulation E error-resolution procedures",
"sourceClass": "regulation",
"accessedAt": "2026-06-01",
"status": "verified",
"access": "sample",
"transformChain": [],
"url": "https://www.consumerfinance.gov/rules-policy/regulations/1005/11"
},
{
"id": "S5",
"label": "CFPB - Regulation Z billing-error resolution",
"sourceClass": "regulation",
"accessedAt": "2026-06-01",
"status": "verified",
"access": "sample",
"transformChain": [],
"url": "https://www.consumerfinance.gov/rules-policy/regulations/1026/13/"
},
{
"id": "S6",
"label": "CFPB - Debt collection examination procedures",
"sourceClass": "regulator",
"accessedAt": "2026-06-01",
"status": "verified",
"access": "sample",
"transformChain": [],
"url": "https://www.consumerfinance.gov/compliance/supervision-examinations/debt-collection-examination-procedures/"
},
{
"id": "S7",
"label": "FTC - Safeguards Rule",
"sourceClass": "regulator",
"accessedAt": "2026-06-01",
"status": "verified",
"access": "sample",
"transformChain": [],
"url": "https://www.ftc.gov/legal-library/browse/rules/safeguards-rule"
},
{
"id": "S8",
"label": "FFIEC - Authentication and access guidance",
"sourceClass": "regulator",
"accessedAt": "2026-06-01",
"status": "verified",
"access": "sample",
"transformChain": [],
"url": "https://www.ffiec.gov/news/press-releases/2021/pr-08-11"
},
{
"id": "S9",
"label": "Federal Reserve - SR 23-4 third-party relationships",
"sourceClass": "regulator",
"accessedAt": "2026-06-01",
"status": "verified",
"access": "sample",
"transformChain": [],
"url": "https://www.federalreserve.gov/supervisionreg/srletters/SR2304.htm"
},
{
"id": "S10",
"label": "CFPB - Service-provider compliance bulletin 2016-02",
"sourceClass": "regulator",
"accessedAt": "2026-06-01",
"status": "verified",
"access": "sample",
"transformChain": [],
"url": "https://www.consumerfinance.gov/compliance/supervisory-guidance/compliance-bulletin-and-policy-guidance-2016-02-service-providers/"
},
{
"id": "S11",
"label": "FCC - AI-generated voice under TCPA ruling",
"sourceClass": "regulator",
"accessedAt": "2026-06-01",
"status": "verified",
"access": "sample",
"transformChain": [],
"url": "https://docs.fcc.gov/public/attachments/FCC-24-17A1_Rcd.pdf"
},
{
"id": "S12",
"label": "NIST - AI Risk Management Framework",
"sourceClass": "standard",
"accessedAt": "2026-06-01",
"status": "verified",
"access": "sample",
"transformChain": [],
"url": "https://www.nist.gov/itl/ai-risk-management-framework"
},
{
"id": "S13",
"label": "European Commission - AI Act overview",
"sourceClass": "regulation",
"accessedAt": "2026-06-01",
"status": "verified",
"access": "sample",
"transformChain": [],
"url": "https://digital-strategy.ec.europa.eu/en/policies/regulatory-framework-ai"
},
{
"id": "S14",
"label": "Sierra - Agent OS",
"sourceClass": "vendor_docs",
"accessedAt": "2026-06-01",
"status": "verified",
"access": "sample",
"transformChain": [],
"url": "https://sierra.ai/product"
},
{
"id": "S15",
"label": "Sierra - Chime customer story",
"sourceClass": "case_study",
"accessedAt": "2026-06-01",
"status": "verified",
"access": "sample",
"transformChain": [],
"url": "https://sierra.ai/customers/chime"
},
{
"id": "S16",
"label": "Decagon - product overview",
"sourceClass": "vendor_docs",
"accessedAt": "2026-06-01",
"status": "verified",
"access": "sample",
"transformChain": [],
"url": "https://decagon.ai/product/overview"
},
{
"id": "S17",
"label": "Decagon - Chime customer story",
"sourceClass": "case_study",
"accessedAt": "2026-06-01",
"status": "verified",
"access": "sample",
"transformChain": [],
"url": "https://decagon.ai/case-studies/chime"
},
{
"id": "S18",
"label": "Decagon - security",
"sourceClass": "privacy_terms",
"accessedAt": "2026-06-01",
"status": "verified",
"access": "sample",
"transformChain": [],
"url": "https://decagon.ai/security"
},
{
"id": "S19",
"label": "Glia - Banking AI Workforce",
"sourceClass": "vendor_docs",
"accessedAt": "2026-06-01",
"status": "verified",
"access": "sample",
"transformChain": [],
"url": "https://www.glia.com/banking-ai-workforce"
},
{
"id": "S20",
"label": "Glia - security",
"sourceClass": "privacy_terms",
"accessedAt": "2026-06-01",
"status": "verified",
"access": "sample",
"transformChain": [],
"url": "https://www.glia.com/security"
},
{
"id": "S21",
"label": "Glia - AI hallucination and prompt-injection guarantee",
"sourceClass": "vendor_press",
"accessedAt": "2026-06-01",
"status": "verified",
"access": "sample",
"transformChain": [],
"url": "https://www.glia.com/news/glia-launches-industry-first-contractual-guarantee-against-ai-hallucinations-and-prompt-injections"
},
{
"id": "S22",
"label": "Glia - Service 1st Federal Credit Union case study",
"sourceClass": "case_study",
"accessedAt": "2026-06-01",
"status": "verified",
"access": "sample",
"transformChain": [],
"url": "https://www.glia.com/case-study/service-1st"
},
{
"id": "S23",
"label": "Interface.ai - banking AI agent platform",
"sourceClass": "vendor_docs",
"accessedAt": "2026-06-01",
"status": "verified",
"access": "sample",
"transformChain": [],
"url": "https://interface.ai/platform/"
},
{
"id": "S24",
"label": "Interface.ai - integrations",
"sourceClass": "vendor_docs",
"accessedAt": "2026-06-01",
"status": "verified",
"access": "sample",
"transformChain": [],
"url": "https://interface.ai/integrations/"
},
{
"id": "S25",
"label": "Interface.ai - Trust Center",
"sourceClass": "privacy_terms",
"accessedAt": "2026-06-01",
"status": "verified",
"access": "sample",
"transformChain": [],
"url": "https://trust.interface.ai/"
},
{
"id": "S26",
"label": "Interface.ai - Dover Federal Credit Union case study",
"sourceClass": "case_study",
"accessedAt": "2026-06-01",
"status": "verified",
"access": "sample",
"transformChain": [],
"url": "https://interface.ai/case-studies/dover-federal-credit-union/"
},
{
"id": "S27",
"label": "Kasisto - KAI platform",
"sourceClass": "vendor_docs",
"accessedAt": "2026-06-01",
"status": "verified",
"access": "sample",
"transformChain": [],
"url": "https://kasisto.com/products/kai-platform/"
},
{
"id": "S28",
"label": "Kasisto - KAI products",
"sourceClass": "vendor_docs",
"accessedAt": "2026-06-01",
"status": "verified",
"access": "sample",
"transformChain": [],
"url": "https://kasisto.com/products/"
},
{
"id": "S29",
"label": "Kasisto - KAI-GPT launch",
"sourceClass": "vendor_press",
"accessedAt": "2026-06-01",
"status": "verified",
"access": "sample",
"transformChain": [],
"url": "https://kasisto.com/press-releases/kasisto-launches-kai-gpt-the-first-banking-industry-specific-large-language-model/"
},
{
"id": "S30",
"label": "Kasisto - Standard Chartered Stacy case study",
"sourceClass": "case_study",
"accessedAt": "2026-06-01",
"status": "verified",
"access": "sample",
"transformChain": [],
"url": "https://5691986.fs1.hubspotusercontent-na1.net/hubfs/5691986/A%20Digital%20Leader-%20KasistoStandard%20Chartered%20Case%20Study-1.pdf"
},
{
"id": "S31",
"label": "Posh AI - agentic AI platform for banking",
"sourceClass": "vendor_docs",
"accessedAt": "2026-06-01",
"status": "verified",
"access": "sample",
"transformChain": [],
"url": "https://www.posh.ai/platform"
},
{
"id": "S32",
"label": "Posh AI - Trust Center",
"sourceClass": "privacy_terms",
"accessedAt": "2026-06-01",
"status": "verified",
"access": "sample",
"transformChain": [],
"url": "https://www.posh.ai/trust"
},
{
"id": "S33",
"label": "Posh AI - Citadel Credit Union customer story",
"sourceClass": "case_study",
"accessedAt": "2026-06-01",
"status": "verified",
"access": "sample",
"transformChain": [],
"url": "https://www.posh.ai/client-stories/citadel"
},
{
"id": "S34",
"label": "Clinc - virtual banking assistant",
"sourceClass": "vendor_docs",
"accessedAt": "2026-06-01",
"status": "verified",
"access": "sample",
"transformChain": [],
"url": "https://clinc.com/vitualbankingassistant/"
},
{
"id": "S35",
"label": "Clinc - platform documentation",
"sourceClass": "api_docs",
"accessedAt": "2026-06-01",
"status": "partial",
"access": "sample",
"transformChain": [],
"url": "https://api.clinc.com/platform/documentation/v2/platform_reference/query_endpoint.html"
},
{
"id": "S36",
"label": "Clinc - Isbank Maxi case study",
"sourceClass": "case_study",
"accessedAt": "2026-06-01",
"status": "partial",
"access": "sample",
"transformChain": [],
"url": "https://clinc.com/maxi/"
},
{
"id": "S37",
"label": "Personetics - Assist",
"sourceClass": "vendor_docs",
"accessedAt": "2026-06-01",
"status": "partial",
"access": "sample",
"transformChain": [],
"url": "https://personetics.com/products/assist/"
},
{
"id": "S38",
"label": "Personetics - SaaS in financial services position paper",
"sourceClass": "vendor_docs",
"accessedAt": "2026-06-01",
"status": "partial",
"access": "sample",
"transformChain": [],
"url": "https://personetics.com/wp-content/uploads/2025/09/Embracing-SaaS-in-Financial-Services-Personetics-Position-Paper-07-25.pdf"
},
{
"id": "S39",
"label": "ElevenLabs - Revolut customer story",
"sourceClass": "case_study",
"accessedAt": "2026-06-01",
"status": "partial",
"access": "sample",
"transformChain": [],
"url": "https://elevenlabs.io/blog/revolut"
},
{
"id": "S40",
"label": "ElevenLabs - agents pricing",
"sourceClass": "pricing_page",
"accessedAt": "2026-06-01",
"status": "partial",
"access": "sample",
"transformChain": [],
"url": "https://elevenlabs.io/pricing/agents"
},
{
"id": "S41",
"label": "ElevenLabs - data residency",
"sourceClass": "privacy_terms",
"accessedAt": "2026-06-01",
"status": "partial",
"access": "sample",
"transformChain": [],
"url": "https://elevenlabs.io/docs/overview/administration/data-residency"
},
{
"id": "S42",
"label": "CFPB - Consumer Complaint Program",
"sourceClass": "regulator",
"accessedAt": "2026-06-01",
"status": "verified",
"access": "sample",
"transformChain": [],
"url": "https://www.consumerfinance.gov/compliance/consumer-complaint-program/"
},
{
"id": "S43",
"label": "Federal Reserve - Interagency information security standards",
"sourceClass": "regulator",
"accessedAt": "2026-06-01",
"status": "verified",
"access": "sample",
"transformChain": [],
"url": "https://www.federalreserve.gov/supervisionreg/interagencyguidelines.htm"
},
{
"id": "S44",
"label": "NCUA - Regulations and guidance for cybersecurity",
"sourceClass": "regulator",
"accessedAt": "2026-06-01",
"status": "verified",
"access": "sample",
"transformChain": [],
"url": "https://ncua.gov/regulation-supervision/regulatory-compliance-resources/cybersecurity-resources/ncuas-regulations-and-guidance"
},
{
"id": "S45",
"label": "EBA - Guidelines on ICT and security risk management",
"sourceClass": "regulator",
"accessedAt": "2026-06-01",
"status": "verified",
"access": "sample",
"transformChain": [],
"url": "https://www.eba.europa.eu/activities/single-rulebook/regulatory-activities/internal-governance/guidelines-ict-and-security-risk-management?version=2019"
},
{
"id": "S46",
"label": "EBA - DORA registers of information preparation",
"sourceClass": "regulator",
"accessedAt": "2026-06-01",
"status": "verified",
"access": "sample",
"transformChain": [],
"url": "https://www.eba.europa.eu/activities/direct-supervision-and-oversight/digital-operational-resilience-act/preparation-dora-application"
},
{
"id": "S47",
"label": "EBA - ICT guideline amendments in context of DORA",
"sourceClass": "regulator",
"accessedAt": "2026-06-01",
"status": "verified",
"access": "sample",
"transformChain": [],
"url": "https://www.eba.europa.eu/publications-and-media/press-releases/eba-amends-its-guidelines-ict-and-security-risk-management-measures-context-dora-application"
},
{
"id": "S48",
"label": "FFIEC - BSA/AML compliance program",
"sourceClass": "regulator",
"accessedAt": "2026-06-01",
"status": "verified",
"access": "sample",
"transformChain": [],
"url": "https://bsaaml.ffiec.gov/manual/AssessingTheBSAAMLComplianceProgram/01"
},
{
"id": "S49",
"label": "FFIEC - Customer Identification Program",
"sourceClass": "regulator",
"accessedAt": "2026-06-01",
"status": "verified",
"access": "sample",
"transformChain": [],
"url": "https://bsaaml.ffiec.gov/manual/AssessingComplianceWithBSARegulatoryRequirements/01"
},
{
"id": "S50",
"label": "FFIEC - Customer Due Diligence",
"sourceClass": "regulator",
"accessedAt": "2026-06-01",
"status": "verified",
"access": "sample",
"transformChain": [],
"url": "https://bsaaml.ffiec.gov/manual/AssessingComplianceWithBSARegulatoryRequirements/02"
},
{
"id": "S51",
"label": "FFIEC - Beneficial Ownership Requirements",
"sourceClass": "regulator",
"accessedAt": "2026-06-01",
"status": "verified",
"access": "sample",
"transformChain": [],
"url": "https://bsaaml.ffiec.gov/manual/AssessingComplianceWithBSARegulatoryRequirements/03"
},
{
"id": "S52",
"label": "FFIEC - Suspicious Activity Reporting",
"sourceClass": "regulator",
"accessedAt": "2026-06-01",
"status": "verified",
"access": "sample",
"transformChain": [],
"url": "https://bsaaml.ffiec.gov/manual/AssessingComplianceWithBSARegulatoryRequirements/04"
},
{
"id": "S53",
"label": "FFIEC - Office of Foreign Assets Control",
"sourceClass": "regulator",
"accessedAt": "2026-06-01",
"status": "verified",
"access": "sample",
"transformChain": [],
"url": "https://bsaaml.ffiec.gov/manual/OfficeOfForeignAssetsControl/01"
},
{
"id": "S54",
"label": "CFPB - AI adverse-action guidance",
"sourceClass": "regulator",
"accessedAt": "2026-06-01",
"status": "verified",
"access": "sample",
"transformChain": [],
"url": "https://www.consumerfinance.gov/about-us/newsroom/cfpb-issues-guidance-on-credit-denials-by-lenders-using-artificial-intelligence/"
},
{
"id": "S55",
"label": "OCC - Model Risk Management revised guidance",
"sourceClass": "regulator",
"accessedAt": "2026-06-01",
"status": "verified",
"access": "sample",
"transformChain": [],
"url": "https://occ.gov/news-issuances/bulletins/2026/bulletin-2026-13.html"
},
{
"id": "S56",
"label": "FDIC, Federal Reserve and OCC - Third-party risk guide for community banks",
"sourceClass": "regulator",
"accessedAt": "2026-06-01",
"status": "verified",
"access": "sample",
"transformChain": [],
"url": "https://www.fdic.gov/news/financial-institution-letters/2024/third-party-risk-management-guide-community-banks"
},
{
"id": "S57",
"label": "U.S. Treasury - Artificial intelligence in financial services",
"sourceClass": "regulator",
"accessedAt": "2026-06-01",
"status": "verified",
"access": "sample",
"transformChain": [],
"url": "https://home.treasury.gov/system/files/136/Artificial-Intelligence-in-Financial-Services.pdf"
},
{
"id": "S58",
"label": "U.S. Treasury - AI-specific cybersecurity risks in financial services",
"sourceClass": "regulator",
"accessedAt": "2026-06-01",
"status": "verified",
"access": "sample",
"transformChain": [],
"url": "https://home.treasury.gov/system/files/136/Managing-Artificial-Intelligence-Specific-Cybersecurity-Risks-In-The-Financial-Services-Sector.pdf"
},
{
"id": "S59",
"label": "Bank of England and FCA - AI in UK financial services 2024",
"sourceClass": "regulator",
"accessedAt": "2026-06-01",
"status": "verified",
"access": "sample",
"transformChain": [],
"url": "https://www.bankofengland.co.uk/report/2024/artificial-intelligence-in-uk-financial-services-2024"
},
{
"id": "S60",
"label": "Basel Committee - Managing AI in banking",
"sourceClass": "regulator",
"accessedAt": "2026-06-01",
"status": "verified",
"access": "sample",
"transformChain": [],
"url": "https://www.bis.org/speeches/sp240417.htm"
},
{
"id": "S61",
"label": "FINRA - Regulatory Notice 24-09 on generative AI",
"sourceClass": "regulator",
"accessedAt": "2026-06-01",
"status": "verified",
"access": "sample",
"transformChain": [],
"url": "https://www.finra.org/rules-guidance/notices/24-09"
},
{
"id": "S62",
"label": "SEC - Regulation S-P amendments",
"sourceClass": "regulator",
"accessedAt": "2026-06-01",
"status": "verified",
"access": "sample",
"transformChain": [],
"url": "https://www.sec.gov/rules-regulations/2024/06/s7-05-23"
},
{
"id": "S63",
"label": "NAIC - Model bulletin on AI systems by insurers",
"sourceClass": "regulator",
"accessedAt": "2026-06-01",
"status": "verified",
"access": "sample",
"transformChain": [],
"url": "https://content.naic.org/sites/default/files/inline-files/2023-12-4%20Model%20Bulletin_Adopted_0.pdf"
},
{
"id": "S64",
"label": "CFPB - Regulation E remittance-transfer error resolution",
"sourceClass": "regulation",
"accessedAt": "2026-06-01",
"status": "verified",
"access": "sample",
"transformChain": [],
"url": "https://www.consumerfinance.gov/rules-policy/regulations/1005/33/"
},
{
"id": "S65",
"label": "CFPB - Regulation X mortgage-servicing error resolution",
"sourceClass": "regulation",
"accessedAt": "2026-06-01",
"status": "verified",
"access": "sample",
"transformChain": [],
"url": "https://www.consumerfinance.gov/rules-policy/regulations/1024/35/"
},
{
"id": "S66",
"label": "CFPB - Regulation X loss-mitigation procedures",
"sourceClass": "regulation",
"accessedAt": "2026-06-01",
"status": "verified",
"access": "sample",
"transformChain": [],
"url": "https://www.consumerfinance.gov/rules-policy/regulations/1024/41/"
},
{
"id": "S67",
"label": "CFPB - Regulation B adverse-action notifications",
"sourceClass": "regulation",
"accessedAt": "2026-06-01",
"status": "verified",
"access": "sample",
"transformChain": [],
"url": "https://www.consumerfinance.gov/rules-policy/regulations/1002/9/"
},
{
"id": "S68",
"label": "FCA - Consumer Duty finalised guidance FG22/5",
"sourceClass": "regulator",
"accessedAt": "2026-06-01",
"status": "verified",
"access": "sample",
"transformChain": [],
"url": "https://www.fca.org.uk/publication/finalised-guidance/fg22-5.pdf"
},
{
"id": "S69",
"label": "European Commission - draft AI Act transparency-obligation guidelines consultation",
"sourceClass": "regulator",
"accessedAt": "2026-06-01",
"status": "verified",
"access": "sample",
"transformChain": [],
"url": "https://digital-strategy.ec.europa.eu/en/news/commission-opens-consultation-draft-guidelines-ai-transparency-obligations"
},
{
"id": "S70",
"label": "SEC - Regulation S-P small-firm compliance outreach",
"sourceClass": "regulator",
"accessedAt": "2026-06-01",
"status": "verified",
"access": "sample",
"transformChain": [],
"url": "https://www.sec.gov/newsroom/meetings-events/compliance-outreach-regulation-s-p-small-firms"
},
{
"id": "S71",
"label": "U.S. Treasury - AI Innovation Series launch",
"sourceClass": "regulator",
"accessedAt": "2026-06-01",
"status": "verified",
"access": "sample",
"transformChain": [],
"url": "https://home.treasury.gov/news/press-releases/sb0421"
},
{
"id": "S72",
"label": "NAIC - AI model bulletin legal adoption map",
"sourceClass": "regulator",
"accessedAt": "2026-06-01",
"status": "verified",
"access": "sample",
"transformChain": [],
"url": "https://content.naic.org/sites/default/files/legal-adoption-map-ai-model-bulletin.pdf"
},
{
"id": "S73",
"label": "European Commission - high-risk AI systems guidelines consultation",
"sourceClass": "regulator",
"accessedAt": "2026-06-01",
"status": "verified",
"access": "sample",
"transformChain": [],
"url": "https://digital-strategy.ec.europa.eu/en/policies/guidelines-ai-high-risk-systems"
}
],
"evidence": [
{
"id": "E1",
"sourceId": "S1",
"claimIds": [
"C1",
"C4"
],
"summary": "The CFPB says financial institutions risk legal obligations, trust and consumer harm when deficient chatbots replace competent support, especially when complex issues need human help.",
"status": "verified",
"confidence": 0.92,
"access": "sample"
},
{
"id": "E2",
"sourceId": "S1",
"claimIds": [
"C1",
"C4",
"C5"
],
"summary": "The CFPB report identifies dispute recognition, inaccurate information, human-access barriers, downtime, impersonation and privacy risks as consumer-finance chatbot failure modes.",
"status": "verified",
"confidence": 0.92,
"access": "sample"
},
{
"id": "E3",
"sourceId": "S2",
"claimIds": [
"C1",
"C4"
],
"summary": "The CFPB blog states that inaccurate customer-service information can become a UDAAP issue and that financial companies have obligations around customer questions, payments and access to money.",
"status": "verified",
"confidence": 0.9,
"access": "sample"
},
{
"id": "E4",
"sourceId": "S3",
"claimIds": [
"C1",
"C4"
],
"summary": "Extracted signal: UDAAP examination procedures cover unfair, deceptive and abusive acts or practices in supervised financial activity. Report inference: financial-institution service AI needs conduct-risk evidence, not only technology controls.",
"status": "verified",
"confidence": 0.87,
"access": "sample"
},
{
"id": "E5",
"sourceId": "S4",
"claimIds": [
"C5"
],
"summary": "Regulation E error-resolution procedures make electronic-fund-transfer error recognition, documentation and response a hard boundary for account and debit-payment service agents.",
"status": "verified",
"confidence": 0.9,
"access": "sample"
},
{
"id": "E6",
"sourceId": "S5",
"claimIds": [
"C5"
],
"summary": "Regulation Z billing-error resolution rules make credit-card and open-end credit disputes a separate route from general account-service automation.",
"status": "verified",
"confidence": 0.9,
"access": "sample"
},
{
"id": "E7",
"sourceId": "S6",
"claimIds": [
"C5"
],
"summary": "CFPB debt-collection procedures cover communications, disputes, complaints, payment processing and harassment or abusive-conduct review, so collections agents need stricter channel and script controls than routine service agents.",
"status": "verified",
"confidence": 0.88,
"access": "sample"
},
{
"id": "E8",
"sourceId": "S7",
"claimIds": [],
"summary": "The FTC Safeguards Rule requires covered financial institutions to maintain safeguards for customer information and to ensure service providers safeguard customer information in their care.",
"status": "verified",
"confidence": 0.9,
"access": "sample"
},
{
"id": "E9",
"sourceId": "S8",
"claimIds": [
"C3"
],
"summary": "FFIEC authentication guidance emphasizes risk assessments, layered security, access controls, customers, employees, service accounts and third-party access to digital banking services.",
"status": "verified",
"confidence": 0.9,
"access": "sample"
},
{
"id": "E10",
"sourceId": "S9",
"claimIds": [],
"summary": "SR 23-4 frames third-party risk management across planning, due diligence, contract negotiation, ongoing monitoring and termination, making AI-agent vendors a lifecycle control rather than a one-off software review.",
"status": "verified",
"confidence": 0.9,
"access": "sample"
},
{
"id": "E11",
"sourceId": "S10",
"claimIds": [],
"summary": "CFPB service-provider guidance expects supervised entities to manage providers in a way that assures compliance with federal consumer financial law and avoids consumer harm.",
"status": "verified",
"confidence": 0.88,
"access": "sample"
},
{
"id": "E12",
"sourceId": "S11",
"claimIds": [],
"summary": "The FCC ruling confirms that TCPA restrictions on artificial or prerecorded voice encompass current AI technologies that generate or resemble human voices, which affects outbound and automated voice routes.",
"status": "verified",
"confidence": 0.9,
"access": "sample"
},
{
"id": "E13",
"sourceId": "S12",
"claimIds": [],
"summary": "Extracted signal: NIST AI RMF supplies governance, mapping, measurement and management categories for AI risk. Report inference: those categories are useful route evidence, but they do not replace financial-services law and supervisor expectations.",
"status": "verified",
"confidence": 0.86,
"access": "sample"
},
{
"id": "E14",
"sourceId": "S13",
"claimIds": [],
"summary": "Extracted signal: The EU AI Act creates transparency duties for certain interactive AI systems and stricter treatment for high-risk uses. Report inference: EU customer-facing banking agents need a separate disclosure and deployment lane.",
"status": "verified",
"confidence": 0.86,
"access": "sample"
},
{
"id": "E15",
"sourceId": "S14",
"claimIds": [
"C2"
],
"summary": "Sierra Agent OS public materials support an enterprise agent platform route with system actions, voice, live assist, observability, simulations, data platform and outcome-oriented deployment claims.",
"status": "verified",
"confidence": 0.86,
"access": "sample"
},
{
"id": "E16",
"sourceId": "S15",
"claimIds": [],
"summary": "Sierra's Chime customer story reports banking support for direct-deposit status, account access, password reset and routine member inquiries, with resolution moving from 50% to more than 70%.",
"status": "verified",
"confidence": 0.82,
"access": "sample"
},
{
"id": "E17",
"sourceId": "S16",
"claimIds": [
"C2"
],
"summary": "Decagon positions its platform around chat, email, voice, AOPs, integrations, testing, QA, analytics, Watchtower and proactive agents, supporting a cross-system agent-platform lane.",
"status": "verified",
"confidence": 0.86,
"access": "sample"
},
{
"id": "E18",
"sourceId": "S17",
"claimIds": [],
"summary": "Decagon's Chime case reports 70% chat and voice resolution, 30 sNPS, lower customer-support costs and more than one million monthly voice calls, but the numbers are vendor-published case evidence rather than an audited benchmark.",
"status": "verified",
"confidence": 0.82,
"access": "sample"
},
{
"id": "E19",
"sourceId": "S18",
"claimIds": [],
"summary": "Decagon security materials describe controls such as RBAC, SSO, short-lived tokens, scoped customer-system access, PII redaction, supervisor checks and audit logs, with contract terms still needed for regulated deployment.",
"status": "verified",
"confidence": 0.84,
"access": "sample"
},
{
"id": "E20",
"sourceId": "S19",
"claimIds": [
"C2"
],
"summary": "Glia's Banking AI Workforce positions the platform around banking-specific AI, voice, digital service, handoff, banking goals and contact-centre workflows.",
"status": "verified",
"confidence": 0.86,
"access": "sample"
},
{
"id": "E21",
"sourceId": "S20",
"claimIds": [],
"summary": "Glia security materials claim SOC 2 Type 2, PCI DSS, HIPAA/HITECH Type 1, CCPA, encryption, penetration testing, PII masking and AWS hosting, which creates a strong but still contract-scoped trust lane.",
"status": "verified",
"confidence": 0.86,
"access": "sample"
},
{
"id": "E22",
"sourceId": "S21",
"claimIds": [
"C4"
],
"summary": "Glia's March 2026 announcement says it will offer a contractual guarantee against AI hallucinations being presented to customers or members and against prompt-injection impact on its Banking AI platform.",
"status": "verified",
"confidence": 0.84,
"access": "sample"
},
{
"id": "E23",
"sourceId": "S22",
"claimIds": [],
"summary": "Glia's Service 1st case reports voice AI answering all calls, handling 37% fully, reducing abandonment by 96%, reducing wait time by 91% and saving 69 agent hours per week.",
"status": "verified",
"confidence": 0.82,
"access": "sample"
},
{
"id": "E24",
"sourceId": "S23",
"claimIds": [
"C2",
"C3"
],
"summary": "Interface.ai's banking AI Agent platform claims a combination of LLM foundation models, rule-based workflows, MFA, access control, data masking and integration to more than 25 core banking systems.",
"status": "verified",
"confidence": 0.87,
"access": "sample"
},
{
"id": "E25",
"sourceId": "S24",
"claimIds": [],
"summary": "Interface.ai integration materials list core, digital-banking, loan-origination, CRM, data-aggregation, credit-scoring and customer-support integration categories, making connector depth a central diligence question.",
"status": "verified",
"confidence": 0.85,
"access": "sample"
},
{
"id": "E26",
"sourceId": "S25",
"claimIds": [],
"summary": "Interface.ai's Trust Center lists GLBA, ISO/IEC 27001:2022, SOC 2 Type 2, CSA STAR, GDPR and related security artefacts, with buyer access needed for exact report scope and evidence.",
"status": "verified",
"confidence": 0.86,
"access": "sample"
},
{
"id": "E27",
"sourceId": "S26",
"claimIds": [],
"summary": "Interface.ai's Dover Federal Credit Union case reports day-one call automation, wait-hour elimination and operating savings for Beau voice-assistant deployment.",
"status": "verified",
"confidence": 0.84,
"access": "sample"
},
{
"id": "E28",
"sourceId": "S27",
"claimIds": [
"C2",
"C4"
],
"summary": "Kasisto KAI platform materials position KAI-GPT, banking-specific agents, AI-plus-human teams and live-chat/contact-centre integration as a purpose-built banking route.",
"status": "verified",
"confidence": 0.86,
"access": "sample"
},
{
"id": "E29",
"sourceId": "S28",
"claimIds": [
"C4"
],
"summary": "Kasisto products materials say KAI Answers can generate replies from internal and public-facing documents for consumer-facing banking agents, so source approval and answer governance matter.",
"status": "verified",
"confidence": 0.84,
"access": "sample"
},
{
"id": "E30",
"sourceId": "S29",
"claimIds": [
"C4"
],
"summary": "Kasisto's KAI-GPT launch positioned a banking-specific LLM route already working with 47 major financial institutions, which supports the vertical-source thesis but remains vendor-published evidence.",
"status": "verified",
"confidence": 0.82,
"access": "sample"
},
{
"id": "E31",
"sourceId": "S30",
"claimIds": [],
"summary": "Kasisto's Standard Chartered case study reports more than 80% containment, 50% live-chat interaction reduction and increased engagement for the Stacy virtual assistant.",
"status": "verified",
"confidence": 0.82,
"access": "sample"
},
{
"id": "E32",
"sourceId": "S31",
"claimIds": [
"C2",
"C3"
],
"summary": "Posh AI's platform page positions the product as a governed agentic banking platform with voice, digital, employee knowledge, operating procedures, secure actions and core/telephony/KMS integrations.",
"status": "verified",
"confidence": 0.86,
"access": "sample"
},
{
"id": "E33",
"sourceId": "S32",
"claimIds": [],
"summary": "Posh's trust materials claim SOC 2 Type II, CSA STAR, auditability, encryption, MFA, monitoring, penetration testing and decision logging, making evidence export and report scope key contract items.",
"status": "verified",
"confidence": 0.85,
"access": "sample"
},
{
"id": "E34",
"sourceId": "S33",
"claimIds": [],
"summary": "Posh's Citadel Credit Union story reports 1.5 million calls, 2 million chats, 21,000 hours saved, Digital Assistant containment and third-party cost savings, all vendor/customer-story evidence.",
"status": "verified",
"confidence": 0.82,
"access": "sample"
},
{
"id": "E35",
"sourceId": "S34",
"claimIds": [
"C2",
"C3"
],
"summary": "Clinc's virtual banking assistant is designed for authenticated banking, account servicing, transactions and voice/chat support, keeping it inside a financial-services vertical lane.",
"status": "verified",
"confidence": 0.84,
"access": "sample"
},
{
"id": "E36",
"sourceId": "S35",
"claimIds": [
"C3"
],
"summary": "Clinc documentation exposes query metadata, session IDs, intent and slot outputs, latency, language/device fields, FAQ filters and business-logic handoff, which are useful audit and integration artefacts.",
"status": "partial",
"confidence": 0.84,
"access": "sample"
},
{
"id": "E37",
"sourceId": "S36",
"claimIds": [],
"summary": "Clinc's Isbank Maxi case reports a production banking assistant handling payments, transfers and personal-finance queries at large scale, with vendor-published usage and transaction metrics.",
"status": "partial",
"confidence": 0.82,
"access": "sample"
},
{
"id": "E38",
"sourceId": "S37",
"claimIds": [
"C2"
],
"summary": "Personetics Assist is positioned as conversational self-service for financial services, using account and transaction context for banking support across digital channels.",
"status": "partial",
"confidence": 0.82,
"access": "sample"
},
{
"id": "E39",
"sourceId": "S38",
"claimIds": [],
"summary": "Personetics SaaS materials describe active-user/product pricing, Azure deployment, single-tenant approach and support/security updates; Assist-specific controls still require buyer diligence.",
"status": "partial",
"confidence": 0.8,
"access": "sample"
},
{
"id": "E40",
"sourceId": "S39",
"claimIds": [],
"summary": "ElevenLabs' Revolut story reports first-line voice support in the UK and Europe with language switching, secure system connection and support for account-specific support, disputes and chargebacks where possible.",
"status": "partial",
"confidence": 0.81,
"access": "sample"
},
{
"id": "E41",
"sourceId": "S40",
"claimIds": [],
"summary": "ElevenLabs Agents publishes self-serve call-minute and concurrency pricing plus enterprise terms, making voice infrastructure cost more transparent than most banking-specialist platforms.",
"status": "partial",
"confidence": 0.84,
"access": "sample"
},
{
"id": "E42",
"sourceId": "S41",
"claimIds": [],
"summary": "ElevenLabs data-residency materials show that voice-agent infrastructure can change data location and integration exposure; optional custom LLMs and webhooks may alter processing paths.",
"status": "partial",
"confidence": 0.82,
"access": "sample"
},
{
"id": "E43",
"sourceId": "S42",
"claimIds": [
"C5"
],
"summary": "The CFPB Consumer Complaint Program says Consumer Response has worked since 2011 to get consumers timely responses from thousands of financial companies, making complaint intake and response a regulator-visible workflow.",
"status": "verified",
"confidence": 0.86,
"access": "sample"
},
{
"id": "E44",
"sourceId": "S4",
"claimIds": [
"C5"
],
"summary": "Regulation E makes documentation and timing part of error-resolution compliance, so an agent that fails to recognise an error can create downstream control failure.",
"status": "verified",
"confidence": 0.88,
"access": "sample"
},
{
"id": "E45",
"sourceId": "S9",
"claimIds": [],
"summary": "Third-party guidance supports contract rights over reporting, audit, incident, termination and ongoing monitoring for AI service vendors.",
"status": "verified",
"confidence": 0.86,
"access": "sample"
},
{
"id": "E46",
"sourceId": "S1",
"claimIds": [],
"summary": "CFPB complaint examples and analysis make false resolution, loops without human access and wrong answers measurable failure classes for private evals.",
"status": "verified",
"confidence": 0.88,
"access": "sample"
},
{
"id": "E47",
"sourceId": "S43",
"claimIds": [],
"summary": "The interagency information-security standards apply banking-sector administrative, technical and physical safeguards to customer information, including security, confidentiality, integrity and disposal expectations.",
"status": "verified",
"confidence": 0.9,
"access": "sample"
},
{
"id": "E48",
"sourceId": "S44",
"claimIds": [],
"summary": "NCUA cybersecurity guidance anchors credit-union safeguarding, service-provider oversight and response planning for unauthorised member-information access.",
"status": "verified",
"confidence": 0.88,
"access": "sample"
},
{
"id": "E49",
"sourceId": "S45",
"claimIds": [],
"summary": "EBA ICT and security-risk guidance adds EU financial-institution expectations for ICT governance, information security, operations, continuity and third-party-provider controls.",
"status": "verified",
"confidence": 0.88,
"access": "sample"
},
{
"id": "E50",
"sourceId": "S46",
"claimIds": [],
"summary": "The EBA says DORA became applicable on 17 January 2025 and requires in-scope financial entities to keep registers of contractual arrangements with ICT third-party service providers for entity, sub-consolidated and consolidated oversight.",
"status": "verified",
"confidence": 0.91,
"access": "sample"
},
{
"id": "E51",
"sourceId": "S47",
"claimIds": [],
"summary": "The EBA narrowed older ICT/security-risk guidelines because DORA introduced harmonised ICT risk-management requirements across financial sectors, making DORA the current EU operational-resilience anchor for covered entities.",
"status": "verified",
"confidence": 0.9,
"access": "sample"
},
{
"id": "E52",
"sourceId": "S48",
"claimIds": [
"C18"
],
"summary": "The FFIEC BSA/AML manual says an adequate programme needs internal controls, independent testing, a BSA compliance officer, training, CIP, CDD and beneficial-ownership procedures, so financial-crime routing cannot be hidden inside generic support automation.",
"status": "verified",
"confidence": 0.91,
"access": "sample"
},
{
"id": "E53",
"sourceId": "S49",
"claimIds": [
"C18"
],
"summary": "The FFIEC CIP section anchors identity-verification duties at account opening and customer relationship creation, making KYC intake a distinct route from ordinary authenticated service.",
"status": "verified",
"confidence": 0.9,
"access": "sample"
},
{
"id": "E54",
"sourceId": "S50",
"claimIds": [
"C18"
],
"summary": "The FFIEC CDD section requires risk-based customer profiles, ongoing monitoring, suspicious-activity support and customer-information updates, so service agents should route material profile changes rather than silently update them.",
"status": "verified",
"confidence": 0.91,
"access": "sample"
},
{
"id": "E55",
"sourceId": "S51",
"claimIds": [
"C18"
],
"summary": "Beneficial-ownership rules require written procedures to identify and verify legal-entity beneficial owners, which makes business-account service and fintech onboarding a separate evidence lane.",
"status": "verified",
"confidence": 0.9,
"access": "sample"
},
{
"id": "E56",
"sourceId": "S52",
"claimIds": [
"C18"
],
"summary": "Suspicious-activity reporting guidance makes detection, evaluation, documentation and filing decisions a financial-crime process; customer-facing agents should preserve facts and avoid clearing, suppressing or discussing SAR-sensitive review.",
"status": "verified",
"confidence": 0.9,
"access": "sample"
},
{
"id": "E57",
"sourceId": "S53",
"claimIds": [
"C18"
],
"summary": "The FFIEC OFAC section makes sanctions compliance a route-specific control for screening and transaction/customer restrictions, not a generic fraud-support label.",
"status": "verified",
"confidence": 0.88,
"access": "sample"
},
{
"id": "E58",
"sourceId": "S54",
"claimIds": [
"C18"
],
"summary": "The CFPB's AI adverse-action guidance says creditors using AI or complex models must provide specific and accurate reasons for credit denials, so credit and lending service agents need approved explanation boundaries.",
"status": "verified",
"confidence": 0.9,
"access": "sample"
},
{
"id": "E59",
"sourceId": "S55",
"claimIds": [
"C18"
],
"summary": "Extracted signal: OCC Bulletin 2026-13 revised model-risk guidance and says agencies plan further work on banks' use of AI, including generative, agentic and AI-based models. Report inference: agent routes still need control evidence where classic model-risk scope is unsettled.",
"status": "verified",
"confidence": 0.9,
"access": "sample"
},
{
"id": "E60",
"sourceId": "S56",
"claimIds": [
"C18"
],
"summary": "Extracted signal: The community-bank third-party guide says reliance on third parties can reduce direct operational control without removing responsibility for safe, sound and compliant activity. Report inference: sponsor-bank, BaaS, processor and outsourced-servicer routes need responsibility evidence.",
"status": "verified",
"confidence": 0.9,
"access": "sample"
},
{
"id": "E61",
"sourceId": "S57",
"claimIds": [
"C18"
],
"summary": "The U.S. Treasury AI financial-services report maps AI uses across underwriting, fraud, AML/CFT, customer service, risk management, cybersecurity and operations, supporting a broad financial-institution coverage map instead of a contact-centre-only view.",
"status": "verified",
"confidence": 0.88,
"access": "sample"
},
{
"id": "E62",
"sourceId": "S58",
"claimIds": [
"C18"
],
"summary": "Treasury's AI cybersecurity report highlights AI-enabled fraud, cybersecurity and operational-resilience risks and notes that smaller institutions often depend on third-party or core-service providers for AI capabilities.",
"status": "verified",
"confidence": 0.87,
"access": "sample"
},
{
"id": "E63",
"sourceId": "S59",
"claimIds": [
"C18"
],
"summary": "Extracted signal: The Bank of England and FCA 2024 survey found 75% of respondent firms already using AI, a third of AI use cases implemented by third parties and critical third-party dependencies as the largest expected increase in systemic risk. Report inference: service-agent approval should include operational-resilience and third-party concentration evidence.",
"status": "verified",
"confidence": 0.88,
"access": "sample"
},
{
"id": "E64",
"sourceId": "S60",
"claimIds": [
"C18"
],
"summary": "Extracted signal: The Basel Committee chair framed AI/ML in banking around governance, privacy, hallucination, bias, third-party and financial-stability concerns. Report inference: service agents should be governed as risk-bearing operating systems, not only efficiency tools.",
"status": "verified",
"confidence": 0.86,
"access": "sample"
},
{
"id": "E65",
"sourceId": "S61",
"claimIds": [
"C18"
],
"summary": "FINRA Notice 24-09 reminds broker-dealers that existing securities-law and FINRA obligations continue to apply when firms use generative AI or LLMs, making wealth and brokerage service routes separate from deposit-service routes.",
"status": "verified",
"confidence": 0.89,
"access": "sample"
},
{
"id": "E66",
"sourceId": "S62",
"claimIds": [
"C18"
],
"summary": "SEC Regulation S-P amendments require covered broker-dealers, investment companies, registered investment advisers, funding portals and transfer agents to maintain incident-response and customer-information safeguards, extending the privacy/security route beyond banks.",
"status": "verified",
"confidence": 0.89,
"access": "sample"
},
{
"id": "E67",
"sourceId": "S63",
"claimIds": [
"C18"
],
"summary": "Extracted signal: The NAIC AI model bulletin sets governance, risk-management, internal-control, audit and third-party-AI expectations for insurers. Report inference: insurance policyholder service should not inherit a banking route without insurance controls.",
"status": "verified",
"confidence": 0.86,
"access": "sample"
},
{
"id": "E68",
"sourceId": "S64",
"claimIds": [
"C5",
"C18"
],
"summary": "Regulation E remittance-transfer rules create separate error-resolution timing, content, investigation and remedy requirements for remittance providers, making money-transfer support a distinct service route.",
"status": "verified",
"confidence": 0.9,
"access": "sample"
},
{
"id": "E69",
"sourceId": "S65",
"claimIds": [
"C5",
"C18"
],
"summary": "Regulation X mortgage-servicing error-resolution rules require servicers to evaluate borrower notices by substance, not only label, so mortgage-service agents need intake and handoff controls beyond general support.",
"status": "verified",
"confidence": 0.9,
"access": "sample"
},
{
"id": "E70",
"sourceId": "S66",
"claimIds": [
"C18"
],
"summary": "Regulation X loss-mitigation procedures make hardship and foreclosure-adjacent mortgage service a separate route that should collect facts and route to a servicer owner rather than improvise outcomes.",
"status": "verified",
"confidence": 0.89,
"access": "sample"
},
{
"id": "E71",
"sourceId": "S67",
"claimIds": [],
"summary": "Regulation B notification rules anchor adverse-action notice obligations and specific-reason controls, so credit-service agents need approved status language and human ownership around denial-adjacent interactions.",
"status": "verified",
"confidence": 0.9,
"access": "sample"
},
{
"id": "E72",
"sourceId": "S68",
"claimIds": [
"C18"
],
"summary": "Extracted signal: FCA Consumer Duty guidance sets expectations around customer outcomes, communications and support. Report inference: UK service agents need outcome evidence as well as channel containment.",
"status": "verified",
"confidence": 0.88,
"access": "sample"
},
{
"id": "E73",
"sourceId": "S69",
"claimIds": [],
"summary": "Extracted signal: On 8 May 2026, the European Commission opened consultation on draft AI Act transparency-obligation guidelines; feedback closed on 3 June 2026 and interaction-notice duties apply from 2 August 2026. Report inference: EU service-agent disclosure evidence may change before production launch.",
"status": "verified",
"confidence": 0.9,
"access": "sample"
},
{
"id": "E74",
"sourceId": "S70",
"claimIds": [],
"summary": "Extracted signal: SEC staff held small-firm Regulation S-P outreach ahead of the 3 June 2026 compliance date, including incident-response expectations, customer notices and examination evidence. Report inference: broker-dealer, adviser, fund and transfer-agent service routes need stronger customer-record and incident-response proof.",
"status": "verified",
"confidence": 0.88,
"access": "sample"
},
{
"id": "E75",
"sourceId": "S71",
"claimIds": [],
"summary": "Extracted signal: Treasury's May 2026 AI Innovation Series says AI is increasingly embedded in core financial-services functions from fraud detection and cybersecurity to credit underwriting and operational risk management. Report inference: governance has moved from experiments to enterprise workflow controls.",
"status": "verified",
"confidence": 0.86,
"access": "sample"
},
{
"id": "E76",
"sourceId": "S72",
"claimIds": [
"C18"
],
"summary": "Extracted signal: NAIC's legal adoption map shows state action and pending state action on the AI model bulletin as of 1 May 2026, including adopted states and insurance-specific regulation or guidance. Report inference: insurance service and claims routes need state-specific AIS governance evidence.",
"status": "verified",
"confidence": 0.87,
"access": "sample"
},
{
"id": "E77",
"sourceId": "S73",
"claimIds": [],
"summary": "Extracted signal: The European Commission consultation on guidelines for high-risk AI systems is open until 23 June 2026. Report inference: EU credit, insurance and profiling-service agents may need refreshed classification evidence before launch.",
"status": "verified",
"confidence": 0.9,
"access": "sample"
}
],
"claims": [
{
"id": "C1",
"statement": "Financial-institution service AI is a trust, identity, complaint, payments, conduct and audit market before it is a support-automation market.",
"confidence": 0.92,
"evidenceIds": [
"E1",
"E2",
"E3",
"E4",
"E61",
"E72"
],
"access": "sample"
},
{
"id": "C2",
"statement": "The useful shortlist starts with service authority: approved answer, authenticated read, guided action, staged write, outbound contact, complaint, dispute or payment boundary.",
"confidence": 0.9,
"evidenceIds": [
"E15",
"E17",
"E20",
"E24",
"E28",
"E32",
"E35",
"E38"
],
"access": "sample"
},
{
"id": "C3",
"statement": "Authentication and access controls should be route-specific because customer, employee, service-account and third-party access expose different failure modes.",
"confidence": 0.9,
"evidenceIds": [
"E9",
"E24",
"E32",
"E35",
"E36"
],
"access": "sample"
},
{
"id": "C4",
"statement": "Approved-answer governance is a conduct control: inaccurate, improvised or unreviewed financial-service answers can become consumer-harm evidence.",
"confidence": 0.91,
"evidenceIds": [
"E1",
"E2",
"E3",
"E4",
"E22",
"E28",
"E29",
"E30"
],
"access": "sample"
},
{
"id": "C5",
"statement": "Disputes, EFT errors, billing errors, remittance errors, mortgage-servicing errors, complaints, collections and payment issues should not be measured as ordinary containment.",
"confidence": 0.91,
"evidenceIds": [
"E2",
"E5",
"E6",
"E7",
"E43",
"E44",
"E68",
"E69"
],
"access": "sample"
},
{
"id": "C18",
"statement": "The coverage map should start with institution type because shared service channels hide different banking, payments, lending, securities and insurance obligations.",
"confidence": 0.89,
"evidenceIds": [
"E52",
"E54",
"E58",
"E60",
"E61",
"E63",
"E65",
"E66",
"E67",
"E68",
"E69",
"E70",
"E72",
"E76"
],
"access": "sample"
}
],
"colophon": {
"checkedAt": "2026-06-01",
"latestRevisionAt": "2026-06-17T00:00:00Z",
"reviewStandard": "expert_human_reviewed",
"kinetics": {
"speed": "fast",
"impact": 0.91,
"momentum": 0.9,
"businessImpact": "Financial-institution service AI is moving from FAQ containment into authenticated routes that can affect access to money, disputes, KYC, fraud, credit, investments, insurance claims, collections and customer trust.",
"rateOfChange": "Near-term EU transparency, high-risk classification, model-risk and insurance-AI signals can change approval evidence faster than annual contact-centre renewals.",
"from": "chatbot deflection and IVR containment",
"toward": "route-approved financial-institution agents with identity, conduct, financial-crime, credit, privacy and audit controls"
},
"watch": [
"23 June 2026: European Commission high-risk AI consultation closes; credit, insurance and profiling-service routes may need different classification evidence [E77].",
"2 August 2026: AI Act transparency obligations apply; EU financial-service agents need disclosure evidence that may split from US service-agent approvals [E73]."
],
"gapAnalysis": "Independent end-to-end financial-institution agent benchmarks linking authentication state, entitlement and service-account scope, institution type, approved source, disputes, complaints, KYC, fraud, payment boundaries, credit and insurance decision support, supervised communications, human baselines, repeat contact and audit export."
},
"body": {
"sections": [
{
"id": "executive-summary",
"title": "Executive summary",
"access": "sample",
"dek": "The agent is safe only where the service route is safe.",
"claimIds": [
"C1",
"C2",
"C4",
"C5",
"C18"
],
"evidenceIds": [
"E1",
"E2",
"E3",
"E5",
"E8",
"E9",
"E10",
"E12",
"E15",
"E16",
"E21",
"E22",
"E23",
"E24",
"E47",
"E48",
"E49",
"E50",
"E51",
"E52",
"E53",
"E58",
"E60",
"E61",
"E63",
"E65",
"E66",
"E67",
"E68",
"E69",
"E71",
"E72",
"E73",
"E74",
"E75",
"E76",
"E77"
],
"blocks": [
{
"id": "executive-summary-narrative",
"evidenceIds": [
"E1",
"E2",
"E3",
"E5",
"E8",
"E9",
"E10",
"E12",
"E15",
"E16",
"E21",
"E22",
"E23",
"E24",
"E47",
"E48",
"E49",
"E50",
"E51",
"E52",
"E53",
"E58",
"E60",
"E61",
"E63",
"E65",
"E66",
"E67",
"E68",
"E69",
"E71",
"E72",
"E73",
"E74",
"E75",
"E76",
"E77"
],
"type": "section_narrative",
"lead": "One financial-service channel can hide many regulated jobs, and each job needs its own owner.",
"body": [
"A balance answer, card dispute, payment correction, KYC prompt, fraud explanation, credit-denial response, insurance claim update and adviser message can all arrive through the same chat, voice or app surface. They do not share one customer duty, source set, system-of-record action or repair owner.",
"The CFPB's chatbot work supplies the public warning label: automated financial service can leave people with wrong answers, unresolved disputes and no human path [E1, E2]. Inaccurate service information can also become a consumer-finance compliance problem [E3]."
],
"detail": [
"Treat vendor case metrics as pattern evidence. They show real deployments and useful workflows [E16, E23], but they rarely expose false resolution, repeat contact, complaint recognition, authentication failure, reviewer minutes or auditability across comparable test sets."
]
},
{
"id": "executive-market-context",
"evidenceIds": [
"E20",
"E24",
"E52",
"E54",
"E65",
"E76"
],
"type": "paragraph",
"text": "Service routes now span banks, credit unions, BaaS programmes, payment firms, lenders, broker-dealers, advisers, insurers and financial-crime teams. KYC/CDD need FFIEC evidence [E52, E54]; securities and insurance routes need supervision and AIS governance [E65, E76].",
"variant": "small"
},
{
"id": "executive-approval-context",
"evidenceIds": [
"E3",
"E5",
"E6",
"E7",
"E9",
"E12",
"E36",
"E44",
"E52",
"E56",
"E58",
"E65",
"E67"
],
"type": "paragraph",
"text": "Service authority is the first approval decision. A branch-hours answer is not the same route as a dispute, payment, fraud summary, credit-denial explanation, adviser draft, claim interaction, borrower call, or system writeback.",
"variant": "small"
},
{
"id": "executive-regulatory-frame",
"evidenceIds": [
"E3",
"E47",
"E48",
"E50",
"E52",
"E63",
"E65",
"E67",
"E72",
"E73"
],
"type": "paragraph",
"text": "Regulatory anchors belong behind that authority split: US consumer-finance and BSA/AML sources [E3, E52], bank and credit-union safeguards [E47, E48], EU AI Act and DORA signals [E50, E73], UK Consumer Duty evidence [E72], and securities and insurance anchors [E65, E67].",
"variant": "small"
},
{
"id": "executive-decision-table",
"title": "The buying decision",
"evidenceIds": [
"E1",
"E2",
"E3",
"E5",
"E8",
"E9",
"E10",
"E12",
"E15",
"E16",
"E21",
"E22",
"E23",
"E24",
"E47",
"E48",
"E49",
"E50",
"E51",
"E52",
"E53",
"E58",
"E60",
"E61",
"E63",
"E65",
"E66",
"E67",
"E68",
"E69",
"E71",
"E72",
"E73",
"E74",
"E75",
"E76",
"E77"
],
"type": "table",
"columns": [
{
"key": "decision",
"label": "Decision",
"align": "left"
},
{
"key": "default",
"label": "Default route",
"align": "left"
},
{
"key": "proof",
"label": "Proof to require",
"align": "left"
}
],
"rows": [
{
"decision": "Answer basic service questions.",
"default": "Approve with approved-content source, versioning and escalation.",
"proof": "Current policy source, answer test set, refusal rule and human fallback."
},
{
"decision": "Use authenticated account context.",
"default": "Approve only by channel, identity state and data class.",
"proof": "FFIEC-style access assessment [E9], MFA state, data masking, session logs and role permissions."
},
{
"decision": "Handle disputes, complaints and errors.",
"default": "Gate tightly; treat as regulated intake, not containment.",
"proof": "Recognition tests for Reg E [E5], Reg Z [E6], complaint language, timing triggers, handoff and records."
},
{
"decision": "Route KYC, fraud, AML or sanctions signals.",
"default": "Assist and preserve facts; do not clear, suppress or discuss financial-crime review.",
"proof": "CIP [E53] and CDD source [E54], risk signal, case ID, owner, escalation reason and approved customer-facing language."
},
{
"decision": "Support remittance or mortgage-servicing interactions.",
"default": "Classify error, payment, hardship and loss-mitigation signals as regulated intake.",
"proof": "Notice content, transfer or loan identifiers, timing rule, servicer/provider owner, borrower/sender communication and handoff record."
},
{
"decision": "Support credit, wealth, brokerage or insurance service.",
"default": "Draft, retrieve and route; block autonomous decisioning or regulated communications until separately approved.",
"proof": "Approved source, licensed-data entitlement, supervised communication record, adverse-action or claim-control owner, and archive export."
},
{
"decision": "Take payment, transfer, collection or outbound action.",
"default": "Human gate unless the route is narrow and reversible.",
"proof": "Consent, payment boundary, script approval, action log, rollback and exception queue."
},
{
"decision": "Buy specialist versus platform route.",
"default": "Prefer banking specialist where core/digital banking controls are hard; prefer platform where workflow spans many systems.",
"proof": "Connector depth, action permissions, audit export, contract remedies and same-stack references."
}
]
},
{
"id": "executive-decision-rule",
"evidenceIds": [
"E1",
"E2",
"E3",
"E9",
"E10",
"E13",
"E19",
"E36",
"E44",
"E46",
"E52",
"E56",
"E58",
"E65",
"E67"
],
"type": "paragraph",
"text": "The executive decision rule is simple: approve narrow routes with named evidence, gate consequential routes with human owners, and block any route that cannot export source, identity, action and outcome proof.",
"variant": "small"
},
{
"id": "cx-companion-link",
"evidenceIds": [],
"type": "report_link",
"reportId": "rpt_ai_customer_experience_workflows_2026_05",
"slug": "ai-customer-service-workflows",
"label": "AI for customer-service workflows",
"dek": "Use the broader CX report for market structure and companion coverage across helpdesk, CCaaS, vertical support, voice infrastructure and QA/control-layer routes.",
"relation": "Related report"
},
{
"id": "pricing-economics-companion-link",
"evidenceIds": [],
"type": "report_link",
"reportId": "rpt_ai_customer_service_pricing_ticket_economics_2026_05",
"slug": "ai-customer-service-pricing-ticket-economics",
"label": "AI customer-service pricing and ticket economics",
"dek": "Use the pricing report when outcome pricing, billing-event audit, retained work and cost per accepted resolution drive the decision.",
"relation": "Related report"
},
{
"id": "route-definition-callout",
"title": "Route definition",
"evidenceIds": [
"E1",
"E2",
"E3",
"E5",
"E8",
"E9",
"E10",
"E12",
"E15",
"E16",
"E21",
"E22",
"E23",
"E24",
"E47",
"E48",
"E49",
"E50",
"E51",
"E52",
"E53",
"E58",
"E60",
"E61",
"E63",
"E65",
"E66",
"E67",
"E68",
"E69",
"E71",
"E72",
"E73",
"E74",
"E75",
"E76",
"E77"
],
"type": "callout",
"tone": "neutral",
"body": "Route means the approved path from work to accepted output: the source set, data boundary, model or vendor channel, tools, reviewer, evidence record and reapproval trigger that make an AI result usable."
}
]
},
{
"id": "financial-institution-coverage-map",
"title": "Financial-institution coverage map",
"access": "sample",
"dek": "The institution type decides which route can be approved.",
"claimIds": [
"C18"
],
"evidenceIds": [
"E1",
"E2",
"E3",
"E4",
"E5",
"E6",
"E7",
"E8",
"E9",
"E10",
"E11",
"E13",
"E14",
"E47",
"E48",
"E49",
"E50",
"E51",
"E52",
"E53",
"E54",
"E55",
"E56",
"E57",
"E58",
"E59",
"E60",
"E61",
"E62",
"E63",
"E64",
"E65",
"E66",
"E67",
"E68",
"E69",
"E70",
"E71",
"E72",
"E73",
"E74",
"E75",
"E76"
],
"blocks": [
{
"id": "financial-institution-coverage-map-narrative",
"evidenceIds": [
"E10",
"E52",
"E53",
"E54",
"E55",
"E56",
"E57",
"E58",
"E60",
"E61",
"E63",
"E65",
"E66",
"E67",
"E68",
"E69",
"E70",
"E71",
"E72",
"E75",
"E76"
],
"type": "section_narrative",
"lead": "Financial institutions share channels. They do not share one approval path.",
"body": [
"A customer may call the same number to ask about a balance, dispute a transfer, update business ownership, report fraud, ask about a credit denial, discuss an investment account, or submit an insurance claim. The agent channel looks the same. The regulated work behind it is different.",
"The useful coverage map therefore starts with institution type and workflow consequence. Treasury's financial-services AI report names uses across underwriting, fraud, AML/CFT, customer service, risk management, cybersecurity and operations [E61]. The Bank of England and FCA survey adds adoption and third-party-dependency evidence across financial firms, and Treasury's 2026 AI Innovation Series points to AI moving into core workflows [E63, E75].",
"The approval route should split when the obligation changes. KYC, CDD, beneficial ownership, sanctions and SAR-sensitive fraud review need financial-crime evidence [E52, E56]. Remittance and mortgage-servicing errors need product-specific intake controls [E68, E69]. Broker-dealer, UK conduct and insurance AI governance cannot be collapsed into generic support containment [E65, E72].",
"Member and customer service remains the surface; the approval map has to reach the regulated work that often sits behind the request."
],
"detail": [
"The boundary matters most in outsourced and embedded finance. A fintech or programme manager may operate the interface, but sponsor banks, processors, issuers, servicers, advisers or insurers still need evidence showing who owns the regulated decision and who can repair the failure [E10, E60]."
]
},
{
"id": "institution-lane-table",
"title": "Institution lane decides the evidence",
"evidenceIds": [
"E1",
"E2",
"E3",
"E4",
"E5",
"E6",
"E7",
"E8",
"E9",
"E10",
"E11",
"E47",
"E48",
"E52",
"E53",
"E54",
"E55",
"E56",
"E57",
"E58",
"E60",
"E61",
"E65",
"E66",
"E67",
"E68",
"E69",
"E70",
"E71",
"E72",
"E76"
],
"type": "table",
"columns": [
{
"key": "institution",
"label": "Institution lane",
"align": "left"
},
{
"key": "service_surface",
"label": "Service surface",
"align": "left"
},
{
"key": "hard_boundary",
"label": "Hard boundary",
"align": "left"
},
{
"key": "evidence",
"label": "Evidence to require",
"align": "left"
}
],
"rows": [
{
"institution": "Retail banks and credit unions.",
"service_surface": "Account service, card status, branch/contact-centre support, disputes and complaints.",
"hard_boundary": "Reg E/Z [E5, E6], complaints, UDAAP [E3, E4], authentication, GLBA/safeguards [E8], NCUA or banking information-security controls [E47, E48].",
"evidence": "Approved answers, identity state, error/dispute intake, human queue, access logs and service-provider evidence."
},
{
"institution": "Digital banks, neobanks and BaaS programmes.",
"service_surface": "Deposit, card, payment, direct-deposit, account-access and programme support.",
"hard_boundary": "Sponsor-bank accountability and third-party dependency [E10, E60], programme-manager role and complaint/dispute handoff.",
"evidence": "Responsibility matrix, ledger owner, case owner, third-party monitoring, termination path and customer-facing remediation process."
},
{
"institution": "Card issuers, payments firms and MSBs.",
"service_surface": "Authorisation status, payment exceptions, chargebacks, remittance help, merchant/cardholder support and fraud alerts.",
"hard_boundary": "Money movement, error rights [E5, E68], sanctions screening, AML/CFT monitoring [E52, E57], fraud queues and payment-data handling.",
"evidence": "Processor status, payee/merchant data, consent, exception queue, screening result, fraud case and reversal or chargeback path."
},
{
"institution": "Consumer lenders, auto finance and mortgage servicers.",
"service_surface": "Application status, document collection, servicing, payment help, hardship and denial-adjacent questions.",
"hard_boundary": "Credit decisioning, adverse-action reasons and fair-lending controls [E58, E71], mortgage-servicing errors [E69] and loss-mitigation handoff [E70].",
"evidence": "Approved status language, document checklist, adverse-action owner, servicing notice/error case, loss-mitigation owner, deadline and human approval."
},
{
"institution": "Wealth, brokerage and advisory firms.",
"service_surface": "Client communications, account service, portfolio Q&A, advisor assist and operational follow-up.",
"hard_boundary": "Recommendation, suitability, fair-and-balanced communications, supervision and books-and-records evidence [E65], and customer-information safeguards [E66].",
"evidence": "Human-send rule, principal/supervisory review where required, archive export, source entitlement and Reg S-P safeguard record."
},
{
"institution": "Asset managers and capital-markets firms.",
"service_surface": "Research assist, client reporting, surveillance support, operational exceptions and data-heavy workflow help.",
"hard_boundary": "Licensed-source rights, market-facing communications, trade-execution boundary and recordkeeping [E65, E66].",
"evidence": "Source rights, entitlement filter, reviewer approval, no-trade-execution boundary, archive record and surveillance case trail."
},
{
"institution": "Insurers, brokers and MGAs.",
"service_surface": "Policy service, first notice of loss, claim status, renewal help, coverage questions and fraud intake.",
"hard_boundary": "Underwriting, pricing, claim denial/payment, policyholder communications and third-party AI governance [E67, E76].",
"evidence": "AIS governance record, coverage-source hierarchy, adjuster review, claim note, fairness test and third-party AI inventory."
},
{
"institution": "Financial-crime and risk operations.",
"service_surface": "KYC refresh, identity mismatch, fraud report, sanctions hit, AML alert and case summarization.",
"hard_boundary": "SAR-sensitive evaluation, tipping-off risk, sanctions override, false-negative risk and investigator accountability.",
"evidence": "Risk signal, source links, case ID, investigator note, escalation reason, customer-facing script and no-clearance control."
}
]
},
{
"id": "institution-lane-autonomy-bridge",
"evidenceIds": [
"E52",
"E54",
"E56",
"E58",
"E60",
"E65",
"E67",
"E68",
"E69",
"E70",
"E71",
"E72",
"E76"
],
"type": "paragraph",
"text": "Lane evidence identifies the obligation. Autonomy depth decides whether the approved path may answer, assist, stage work or stay blocked.",
"variant": "small"
},
{
"id": "coverage-depth-table",
"title": "Autonomy depends on decision authority",
"evidenceIds": [
"E52",
"E54",
"E56",
"E58",
"E59",
"E60",
"E61",
"E63",
"E65",
"E66",
"E67",
"E68",
"E69",
"E70",
"E71",
"E72"
],
"type": "table",
"columns": [
{
"key": "depth",
"label": "Depth",
"align": "left"
},
{
"key": "where_it_fits",
"label": "Where it fits",
"align": "left"
},
{
"key": "buyer_move",
"label": "Buyer move",
"align": "left"
}
],
"rows": [
{
"depth": "Full route approval.",
"where_it_fits": "High-volume, customer-facing workflows with clear source, identity, handoff and record requirements.",
"buyer_move": "Approve by institution lane, not by vendor brand or channel."
},
{
"depth": "Assist-only or staged action.",
"where_it_fits": "Credit, fraud, AML, claims, brokerage, advisory and market-facing work where the agent can prepare evidence but not decide.",
"buyer_move": "Require human-send, human-clearance or human-approval gates with exportable records."
},
{
"depth": "Blocked until separately governed.",
"where_it_fits": "Credit decisioning, adverse-action generation, claim denial/payment, trade execution, AML/fraud clearance and unsupervised investment recommendations.",
"buyer_move": "Create a separate regulated route with legal, compliance, model-risk and audit sign-off before any autonomous action."
},
{
"depth": "Companion-report territory.",
"where_it_fits": "Model selection, financial-file parsing, capital-markets research, finance operations and broad CX platform economics.",
"buyer_move": "Use this frame for service authority; use companion reports for model, file, market-data and cost details."
}
]
}
]
},
{
"id": "service-authority-register",
"title": "Institution authority register",
"access": "sample",
"dek": "Approve the action, not the assistant.",
"claimIds": [
"C1",
"C2",
"C3",
"C5",
"C18"
],
"evidenceIds": [
"E1",
"E2",
"E5",
"E6",
"E7",
"E9",
"E10",
"E11",
"E12",
"E13",
"E19",
"E21",
"E24",
"E25",
"E33",
"E36",
"E44",
"E45",
"E47",
"E48",
"E49",
"E50",
"E51",
"E52",
"E54",
"E56",
"E58",
"E68",
"E69",
"E70",
"E71"
],
"blocks": [
{
"id": "service-authority-register-narrative",
"evidenceIds": [
"E1",
"E2",
"E5",
"E6",
"E7",
"E9",
"E12",
"E24",
"E36",
"E44",
"E52",
"E54",
"E56",
"E58",
"E65",
"E66",
"E67",
"E68",
"E69",
"E70",
"E71"
],
"type": "section_narrative",
"lead": "An institution-grade agent needs a route register that says what the agent may know, say, change and escalate.",
"body": [
"The register is the signature proof because it turns a vendor demo into an approval artefact. It separates information, authentication, regulated intake, money movement, KYC, fraud, credit, investments, insurance, collections and outbound voice before the shortlist starts.",
"Most support automation dashboards reward containment. Financial service needs a different first page: which service route is approved, what authority it carries, what evidence it must retain, and when it must stop.",
"The same vendor may be approved for public-product questions, conditional for authenticated balance support, assist-only for KYC refresh, and blocked for dispute closure, SAR-sensitive clearance, claim denial, credit decisioning or investment recommendations. The route, not the logo, carries the risk."
],
"detail": [
"A service-authority register should be owned jointly by the customer channel, product owner, compliance, financial-crime, risk, security, legal, vendor management and the system-of-record owner. If no owner can sign one row, that row is not production-ready."
]
},
{
"id": "service-authority-table",
"title": "Authority decides what can launch",
"evidenceIds": [
"E1",
"E2",
"E5",
"E6",
"E7",
"E9",
"E12",
"E24",
"E36",
"E44",
"E52",
"E53",
"E54",
"E55",
"E56",
"E57",
"E58",
"E65",
"E66",
"E67",
"E68",
"E69",
"E70",
"E71"
],
"type": "table",
"columns": [
{
"key": "route",
"label": "Route",
"align": "left"
},
{
"key": "examples",
"label": "Examples",
"align": "left"
},
{
"key": "default_gate",
"label": "Default gate",
"align": "left"
},
{
"key": "record",
"label": "Record to keep",
"align": "left"
}
],
"rows": [
{
"route": "Public answer",
"examples": "Hours, locations, product FAQs, fee schedule explanations.",
"default_gate": "Approved-answer content and versioning.",
"record": "Source article, answer version, channel, transcript and escalation flag."
},
{
"route": "Authenticated read",
"examples": "Balance, transaction status, direct-deposit status, card state.",
"default_gate": "Identity state, session scope and data-class policy.",
"record": "Authentication event, data source, fields shown, masking and user session."
},
{
"route": "Guided service action",
"examples": "Password reset, card lock, branch appointment, document upload.",
"default_gate": "User confirmation and narrow tool allowlist.",
"record": "Intent, confirmation, tool call, result, reversal path and handoff."
},
{
"route": "Dispute, complaint or error intake",
"examples": "Unauthorised EFT, credit-card billing error, fee complaint, fraud claim.",
"default_gate": "Regulated-intake recognition and human queue.",
"record": "Customer words, product, date, amount, regulatory trigger, deadline and owner."
},
{
"route": "KYC, CDD or beneficial-ownership update",
"examples": "New business owner, changed address, expected-activity change, document gap, identity mismatch.",
"default_gate": "Financial-crime/KYC owner review and event-driven risk-profile handling.",
"record": "Customer information, document source, risk signal, ownership/control data, reviewer and update rationale."
},
{
"route": "Fraud, AML or sanctions signal",
"examples": "Account takeover, suspicious transaction, sanctions hit, unusual activity, payment scam.",
"default_gate": "Assist-only investigation route; no autonomous clearance, suppression or SAR-sensitive discussion.",
"record": "Risk signal, source links, case ID, investigator queue, approved language and escalation reason."
},
{
"route": "Credit or lending support",
"examples": "Application status, document collection, hardship, credit-line question, adverse-action-adjacent question.",
"default_gate": "Approved credit-policy language and human owner for eligibility, denial and exception decisions.",
"record": "Product, application/loan ID, document checklist, source reason, adverse-action owner and handoff."
},
{
"route": "Mortgage servicing or loss mitigation",
"examples": "Payment application question, escrow issue, servicing error notice, hardship request, loss-mitigation status.",
"default_gate": "Servicer owner review, error-intake recognition and approved borrower communication.",
"record": "Borrower words, loan ID, notice/error category, timing trigger, loss-mitigation owner and communication log."
},
{
"route": "Wealth, brokerage or advisory service",
"examples": "Advisor assist, portfolio Q&A, account service, client message draft, transfer instruction.",
"default_gate": "Human-send, supervision, source entitlement and books-and-records archive.",
"record": "Source data, adviser/reviewer, communication version, archive ID, account authority and restriction."
},
{
"route": "Insurance service or claims intake",
"examples": "Policy question, first notice of loss, claim status, coverage question, renewal support.",
"default_gate": "Approved coverage source, adjuster/underwriter review and no autonomous denial or payment.",
"record": "Policy/claim ID, source provision, customer statement, adjuster owner, decision boundary and audit note."
},
{
"route": "Payment or transfer",
"examples": "Loan payment, card payment, ACH transfer, remittance transfer, payment plan.",
"default_gate": "Explicit customer confirmation and high-risk human gate.",
"record": "Payee, amount, account, consent, confirmation, processor status and exception."
},
{
"route": "Collections and hardship",
"examples": "Payment reminder, hardship conversation, promise to pay, cease request.",
"default_gate": "Approved script, channel limits and collector review.",
"record": "Consent basis, script version, contact time, consumer request, promise and escalation."
},
{
"route": "Outbound AI voice",
"examples": "Payment reminder, fraud alert, service callback, appointment reminder.",
"default_gate": "TCPA/consent, disclosure and suppression-list approval.",
"record": "Consent source, called number, disclosure script, recording notice, opt-out and outcome."
}
]
},
{
"id": "authority-layer-bridge",
"evidenceIds": [
"E9",
"E10",
"E13",
"E19",
"E24",
"E36",
"E47",
"E50",
"E51",
"E52",
"E56",
"E68",
"E69",
"E70"
],
"type": "paragraph",
"text": "Route authority says what can launch. Layered authority names the owner, permission and system evidence that keep the approved path from creeping after launch.",
"variant": "small"
},
{
"id": "layered-authority-map",
"title": "Layered authority blocks route creep",
"evidenceIds": [
"E5",
"E6",
"E7",
"E9",
"E10",
"E13",
"E19",
"E21",
"E24",
"E25",
"E33",
"E36",
"E44",
"E45",
"E47",
"E48",
"E49",
"E50",
"E51",
"E52",
"E56",
"E58",
"E68",
"E69",
"E70",
"E71"
],
"type": "table",
"columns": [
{
"key": "authority_layer",
"label": "Authority layer",
"align": "left"
},
{
"key": "system_owner",
"label": "System or owner",
"align": "left"
},
{
"key": "agent_may_do",
"label": "Agent may do",
"align": "left"
},
{
"key": "gate_evidence",
"label": "Gate or evidence",
"align": "left"
},
{
"key": "stop_condition",
"label": "Stop condition",
"align": "left"
}
],
"rows": [
{
"authority_layer": "Customer or member identity and authentication",
"system_owner": "IAM, digital banking and contact-centre authentication.",
"agent_may_do": "Match the interaction to the authenticated customer or member, channel, device and caller state.",
"gate_evidence": "MFA/session state, voice or digital step-up, device/caller risk and failed-authentication log.",
"stop_condition": "Identity is partial, failed, stale or inconsistent with the requested action."
},
{
"authority_layer": "Entitlements, RBAC and service account",
"system_owner": "IAM, cybersecurity, operations role owners and API owners.",
"agent_may_do": "Use only the customer authority, employee role and AI service-account scope assigned to the approved route.",
"gate_evidence": "RBAC matrix, tool allowlist, maker/checker review, privileged-action exception and revocation test.",
"stop_condition": "Shared super-user access, missing revocation or a reviewer approving their own route change."
},
{
"authority_layer": "Core banking, ledger or policy system",
"system_owner": "Core banking, payments, servicing, claims or IT platform owner.",
"agent_may_do": "Read allowed fields or stage reversible changes under the customer's identity and product state.",
"gate_evidence": "Field map, source record, tool payload, idempotency check, approval and rollback path.",
"stop_condition": "Ledger-impacting writeback, fee reversal or account-status change lacks approved route authority."
},
{
"authority_layer": "CRM and case",
"system_owner": "CRM, complaint operations and customer-service case owner.",
"agent_may_do": "Create or update the approved case type, assign the owner and attach the transcript.",
"gate_evidence": "Case payload, routing rule, queue, deadline, owner and closure restriction.",
"stop_condition": "The agent closes, denies or reclassifies a complaint, dispute, error or AI incident."
},
{
"authority_layer": "Contact centre and voice",
"system_owner": "Contact-centre operations, telephony owner and voice compliance.",
"agent_may_do": "Triage, authenticate, transfer, summarize and use approved scripts for narrow voice routes.",
"gate_evidence": "Disclosure, consent, recording notice, suppression list, transfer QA and call outcome.",
"stop_condition": "Disclosure, consent, authentication or transfer proof is missing."
},
{
"authority_layer": "Payments and cards",
"system_owner": "Payments, cards, processor and digital-banking payments owner.",
"agent_may_do": "Explain status, lock or stage a card/payment action where the approved route allows it.",
"gate_evidence": "Payee, amount, account, card state, processor status, customer confirmation and exception queue.",
"stop_condition": "Money moves, payee changes or card/account state changes without route approval."
},
{
"authority_layer": "Disputes, remittances and servicing errors",
"system_owner": "Dispute operations, complaint owner and product compliance.",
"agent_may_do": "Recognise intake, preserve the customer's words, classify product route and start handoff.",
"gate_evidence": "Product, amount, date, deadline, communication history, owner and customer-facing template.",
"stop_condition": "Autonomous denial, closure, discouragement, remedy selection or deadline handling."
},
{
"authority_layer": "Collections",
"system_owner": "Collections operations, servicing compliance and contact-strategy owner.",
"agent_may_do": "Record requests, use approved language, route hardship and schedule human review.",
"gate_evidence": "Script version, contact time, consent basis, restriction status, promise and supervisor review.",
"stop_condition": "Cease request is missed, contact repeats after restriction or negotiation runs without review."
},
{
"authority_layer": "Fraud, AML and KYC routing",
"system_owner": "Fraud operations, financial-crime/KYC owner and security.",
"agent_may_do": "Route suspicious identity, transaction or account-access signals to the owner and preserve context.",
"gate_evidence": "Risk signal, transcript, case ID, queue owner and approved customer-facing language.",
"stop_condition": "The agent clears, suppresses or discusses financial-crime review instead of routing it."
},
{
"authority_layer": "Lending, mortgage servicing and deposit servicing",
"system_owner": "Loan servicing, mortgage servicing, deposit product owner and customer operations.",
"agent_may_do": "Explain approved status, request documents and stage servicing tasks.",
"gate_evidence": "Product policy, account/loan state, approved language, case owner and human handoff.",
"stop_condition": "Autonomous credit decision, eligibility promise, fee refusal, loss-mitigation outcome or hardship outcome."
},
{
"authority_layer": "Audit, model risk and compliance",
"system_owner": "Compliance, model risk, internal audit and operational risk.",
"agent_may_do": "Produce route evidence, eval results, exception queues and incident traces.",
"gate_evidence": "Route approval, test result, model/policy version, reviewer, override and audit export.",
"stop_condition": "No trace can reconstruct source, identity state, action, reviewer and final outcome."
},
{
"authority_layer": "Vendor contract authority",
"system_owner": "Legal, procurement, vendor risk and ICT third-party owner.",
"agent_may_do": "Use vendor functions only inside approved product, data, region, subprocessor and support scope.",
"gate_evidence": "Order form, DPA, security artefacts, subprocessor path, audit right and exit evidence.",
"stop_condition": "Workflow, model route, subprocessor or audit-export need sits outside contract authority."
}
]
},
{
"id": "authority-autonomy-bridge",
"evidenceIds": [
"E1",
"E2",
"E5",
"E6",
"E7",
"E9",
"E12",
"E44"
],
"type": "paragraph",
"text": "Once the authorities and records are named, the buyer can make the harder decision: where the agent may answer, where it may prepare work and where autonomy should stop.",
"variant": "small"
},
{
"id": "service-authority-matrix",
"title": "Autonomy limits by workload",
"evidenceIds": [
"E2",
"E5",
"E6",
"E7",
"E9",
"E12",
"E44"
],
"type": "table",
"columns": [
{
"key": "workload",
"label": "Workload",
"align": "left"
},
{
"key": "answer",
"label": "Answer",
"align": "left"
},
{
"key": "read",
"label": "Read",
"align": "left"
},
{
"key": "stage",
"label": "Stage action",
"align": "left"
},
{
"key": "execute",
"label": "Execute",
"align": "left"
},
{
"key": "rule",
"label": "Rule",
"align": "left"
}
],
"rows": [
{
"workload": "Routine service.",
"answer": "Allow approved content.",
"read": "Allow masked data by identity state.",
"stage": "Gate with customer confirmation.",
"execute": "Limit to reversible actions.",
"rule": "Approve only where source and handoff are controlled."
},
{
"workload": "Disputes and complaints.",
"answer": "Allow rights explanation from approved source.",
"read": "Gate product or account pull.",
"stage": "Gate case opening and deadlines.",
"execute": "Block autonomous denial or closure.",
"rule": "Recognition and routing matter more than containment."
},
{
"workload": "Payments and transfers.",
"answer": "Allow option explanation.",
"read": "Gate account and status verification.",
"stage": "Gate prepared action.",
"execute": "Block unapproved movement.",
"rule": "Money movement needs confirmation, evidence and rollback."
},
{
"workload": "Collections and outbound.",
"answer": "Gate to approved script.",
"read": "Gate account and contact restrictions.",
"stage": "Gate promise or plan proposal.",
"execute": "Block unsupervised pressure.",
"rule": "Channel, timing, consent and tone are part of compliance."
}
]
},
{
"id": "authority-provider-economics-bridge",
"evidenceIds": [
"E16",
"E18",
"E23",
"E27",
"E34",
"E40",
"E41",
"E44",
"E46",
"E60",
"E75"
],
"type": "paragraph",
"text": "Provider lanes come before economics because the denominator changes with ownership. A banking specialist, agent platform, voice layer, fraud case layer or archive layer leaves a different mix of accepted output, retained human work and audit export with the institution.",
"variant": "small"
},
{
"id": "service-authority-register-section-callout",
"title": "Do not average away the regulated case",
"evidenceIds": [
"E1",
"E2",
"E5",
"E6",
"E7",
"E9",
"E10",
"E11",
"E12",
"E13",
"E19",
"E21",
"E24",
"E25",
"E33",
"E36",
"E44",
"E45",
"E47",
"E48",
"E49",
"E50",
"E51",
"E52",
"E54",
"E56",
"E58",
"E68",
"E69",
"E70",
"E71"
],
"type": "callout",
"tone": "info",
"body": "A high containment rate is weak comfort if the missed case was an EFT error, a billing dispute, a sanctions hit, an adverse-action explanation, a claim denial, or a supervised communication."
}
]
},
{
"id": "provider-lanes",
"title": "Provider lanes",
"access": "locked",
"dek": "The market splits by control surface.",
"claimIds": [],
"evidenceIds": [],
"blocks": [
{
"id": "provider-lanes-locked",
"title": "Locked section",
"evidenceIds": [],
"type": "callout",
"tone": "locked",
"body": "Purchase or subscribe to unlock this section."
}
]
},
{
"id": "accepted-output-economics",
"title": "Accepted-output economics",
"access": "locked",
"dek": "The invoice is only useful after the service outcome survives.",
"claimIds": [],
"evidenceIds": [],
"blocks": [
{
"id": "accepted-output-economics-locked",
"title": "Locked section",
"evidenceIds": [],
"type": "callout",
"tone": "locked",
"body": "Purchase or subscribe to unlock this section."
}
]
},
{
"id": "workflow-routing",
"title": "Financial workflow routing",
"access": "locked",
"dek": "Service work is not one workload.",
"claimIds": [],
"evidenceIds": [],
"blocks": [
{
"id": "workflow-routing-locked",
"title": "Locked section",
"evidenceIds": [],
"type": "callout",
"tone": "locked",
"body": "Purchase or subscribe to unlock this section."
}
]
},
{
"id": "identity-data-boundary",
"title": "Identity, entitlement and data boundary",
"access": "locked",
"dek": "Every service route starts with who the agent thinks it is serving.",
"claimIds": [],
"evidenceIds": [],
"blocks": [
{
"id": "identity-data-boundary-locked",
"title": "Locked section",
"evidenceIds": [],
"type": "callout",
"tone": "locked",
"body": "Purchase or subscribe to unlock this section."
}
]
},
{
"id": "approved-answer-controls",
"title": "Regulated communication controls",
"access": "locked",
"dek": "Conduct risk lives in the answer path.",
"claimIds": [],
"evidenceIds": [],
"blocks": [
{
"id": "approved-answer-controls-locked",
"title": "Locked section",
"evidenceIds": [],
"type": "callout",
"tone": "locked",
"body": "Purchase or subscribe to unlock this section."
}
]
},
{
"id": "core-digital-integrations",
"title": "Systems-of-record integrations",
"access": "locked",
"dek": "Integration logos are not action rights.",
"claimIds": [],
"evidenceIds": [],
"blocks": [
{
"id": "core-digital-integrations-locked",
"title": "Locked section",
"evidenceIds": [],
"type": "callout",
"tone": "locked",
"body": "Purchase or subscribe to unlock this section."
}
]
},
{
"id": "voice-outbound-collections",
"title": "Voice, outbound and collections",
"access": "locked",
"dek": "Phone automation is where customer service, consent and pressure meet.",
"claimIds": [],
"evidenceIds": [],
"blocks": [
{
"id": "voice-outbound-collections-locked",
"title": "Locked section",
"evidenceIds": [],
"type": "callout",
"tone": "locked",
"body": "Purchase or subscribe to unlock this section."
}
]
},
{
"id": "evidence-case-studies",
"title": "Evidence and case studies",
"access": "locked",
"dek": "Case studies show workflow shape, not a universal automation ceiling.",
"claimIds": [],
"evidenceIds": [],
"blocks": [
{
"id": "evidence-case-studies-locked",
"title": "Locked section",
"evidenceIds": [],
"type": "callout",
"tone": "locked",
"body": "Purchase or subscribe to unlock this section."
}
]
},
{
"id": "private-evals-auditability",
"title": "Private evals and auditability",
"access": "locked",
"dek": "The eval should test the institution's duty, not the agent's charm.",
"claimIds": [],
"evidenceIds": [],
"blocks": [
{
"id": "private-evals-auditability-locked",
"title": "Locked section",
"evidenceIds": [],
"type": "callout",
"tone": "locked",
"body": "Purchase or subscribe to unlock this section."
}
]
},
{
"id": "deployment-contracts",
"title": "Deployment and contract risk",
"access": "locked",
"dek": "An institution-grade contract should make failure inspectable.",
"claimIds": [],
"evidenceIds": [],
"blocks": [
{
"id": "deployment-contracts-locked",
"title": "Locked section",
"evidenceIds": [],
"type": "callout",
"tone": "locked",
"body": "Purchase or subscribe to unlock this section."
}
]
},
{
"id": "jurisdiction-controls",
"title": "Jurisdiction controls",
"access": "locked",
"dek": "US/EU/UK-led service agents need local approval files.",
"claimIds": [],
"evidenceIds": [],
"blocks": [
{
"id": "jurisdiction-controls-locked",
"title": "Locked section",
"evidenceIds": [],
"type": "callout",
"tone": "locked",
"body": "Purchase or subscribe to unlock this section."
}
]
},
{
"id": "implementation-plan",
"title": "Implementation plan",
"access": "locked",
"dek": "Ninety days should produce an approval artefact, not a pilot anecdote.",
"claimIds": [],
"evidenceIds": [],
"blocks": [
{
"id": "implementation-plan-locked",
"title": "Locked section",
"evidenceIds": [],
"type": "callout",
"tone": "locked",
"body": "Purchase or subscribe to unlock this section."
}
]
},
{
"id": "known-unknowns",
"title": "Known unknowns",
"access": "locked",
"dek": "The missing evidence becomes the diligence brief.",
"claimIds": [],
"evidenceIds": [],
"blocks": [
{
"id": "known-unknowns-locked",
"title": "Locked section",
"evidenceIds": [],
"type": "callout",
"tone": "locked",
"body": "Purchase or subscribe to unlock this section."
}
]
},
{
"id": "vendor-directory",
"title": "Vendor directory",
"access": "locked",
"dek": "Use the directory after selecting the service route.",
"claimIds": [],
"evidenceIds": [],
"blocks": [
{
"id": "vendor-directory-locked",
"title": "Locked section",
"evidenceIds": [],
"type": "callout",
"tone": "locked",
"body": "Purchase or subscribe to unlock this section."
}
]
},
{
"id": "source-ledger",
"title": "Source ledger",
"access": "locked",
"dek": "Primary regulator, standard, vendor, case-study, pricing and trust sources behind the route analysis.",
"claimIds": [],
"evidenceIds": [],
"blocks": [
{
"id": "source-ledger-locked",
"title": "Locked section",
"evidenceIds": [],
"type": "callout",
"tone": "locked",
"body": "Purchase or subscribe to unlock this section."
}
]
},
{
"id": "glossary",
"title": "Glossary",
"access": "sample",
"claimIds": [],
"evidenceIds": [],
"blocks": [
{
"id": "glossary-table",
"title": "Terms used in this report",
"caption": "Definitions are reader aids, not separate evidence claims.",
"evidenceIds": [],
"type": "table",
"columns": [
{
"key": "term",
"label": "Term",
"align": "left"
},
{
"key": "meaning",
"label": "Plain meaning",
"align": "left"
}
],
"rows": [
{
"term": "Route",
"meaning": "The approved path from work to accepted output: source set, data boundary, model or vendor channel, tools, reviewer, evidence record and reapproval trigger."
},
{
"term": "API",
"meaning": "Application programming interface: a governed software connection used by an app or workflow to call a model or service."
},
{
"term": "LLM",
"meaning": "Large language model: a model trained to read and generate text-like tokens."
},
{
"term": "Token",
"meaning": "A chunk of text used for billing, context windows and rate limits. It is usually smaller than a word, not a page count."
}
]
}
]
}
]
}
},
"gap_analysis": "Independent end-to-end financial-institution agent benchmarks linking authentication state, entitlement and service-account scope, institution type, approved source, disputes, complaints, KYC, fraud, payment boundaries, credit and insurance decision support, supervised communications, human baselines, repeat contact and audit export.",
"kinetics": {
"speed": "fast",
"impact": 0.91,
"momentum": 0.9,
"businessImpact": "Financial-institution service AI is moving from FAQ containment into authenticated routes that can affect access to money, disputes, KYC, fraud, credit, investments, insurance claims, collections and customer trust.",
"rateOfChange": "Near-term EU transparency, high-risk classification, model-risk and insurance-AI signals can change approval evidence faster than annual contact-centre renewals.",
"from": "chatbot deflection and IVR containment",
"toward": "route-approved financial-institution agents with identity, conduct, financial-crime, credit, privacy and audit controls"
},
"watch_signals": [
"23 June 2026: European Commission high-risk AI consultation closes; credit, insurance and profiling-service routes may need different classification evidence [E77].",
"2 August 2026: AI Act transparency obligations apply; EU financial-service agents need disclosure evidence that may split from US service-agent approvals [E73]."
],
"sample_findings": [
"Financial-institution service AI should be approved by institution lane and service authority, not by generic chatbot capability."
]
}
```