report.object
AI agents for legal workflows
Expert-reviewed from public records, platform data, primary sources, and private research beyond AI search. Use /api/v1/reports/RPT_01KSHN11BNS8GSRWBRSNY2N409/sample before purchase, then download PDF, JSON, or Markdown after checkout.
page.markdown
AI agents for legal workflows.md
# AI agents for legal workflows
Legal AI agents should be bought as governed workflows: what they may read, draft, send, file, sign, remember, escalate and prove after lawyer review.
## Object
| field | value |
| ---------------- | ------------------------------ |
| report_id | RPT_01KSHN11BNS8GSRWBRSNY2N409 |
| schema_version | 2026-05-28 |
| last_updated_at | 2026-05-31 |
| category | AGENTS |
| sector | Legal services |
| region | GLOBAL |
| confidence_score | 0.89 |
| source_count | 105 |
| evidence_count | 118 |
## Endpoints
| action | endpoint |
| ------------ | --------------------------------------------------------- |
| summary | /api/v1/reports/RPT_01KSHN11BNS8GSRWBRSNY2N409 |
| sample | /api/v1/reports/RPT_01KSHN11BNS8GSRWBRSNY2N409/sample |
| schema | /api/v1/reports/RPT_01KSHN11BNS8GSRWBRSNY2N409/schema |
| quote | /api/v1/reports/RPT_01KSHN11BNS8GSRWBRSNY2N409/quote |
| purchase | /checkout/RPT_01KSHN11BNS8GSRWBRSNY2N409?view=agent |
| versions | /api/v1/reports/RPT_01KSHN11BNS8GSRWBRSNY2N409/versions |
| changelog | /api/v1/reports/RPT_01KSHN11BNS8GSRWBRSNY2N409/changelog |
| monitor | /api/v1/monitors?report_id=RPT_01KSHN11BNS8GSRWBRSNY2N409 |
| mcp_resource | colophon://reports/RPT_01KSHN11BNS8GSRWBRSNY2N409/sample |
## Coverage
| field | value |
| -------------- | ------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------- |
| category | AGENTS |
| sector | Legal services |
| region | GLOBAL |
| topics | legal AI agents, legal AI providers, legal procurement, agent governance, legal retrieval, document intelligence, private evals, cross-border AI, AI regulation, legal AI litigation, AI politics |
| source_classes | official_model_docs, privacy_terms, legal_platform_docs, industry_survey, agent_benchmark, professional_benchmark, retrieval_benchmark, legal_benchmark, legal_ai_analysis, bar_guidance, regulator, document_ai_docs, knowledge_platform_docs, legal_risk_database, job_posting, package_telemetry, issue_tracker, status_page, procurement_record, regulatory_registry, court_docket, archived_web, security_advisory, domain_registry, certificate_transparency, grant_database, civic_complaint, lobbying_disclosure, environmental_compliance, bid_protest, municipal_disclosure, workplace_safety, takedown_notice, bug_bounty, standards_body, expert_profile, evaluation_framework, case_study, demo_video, expert_commentary, industry_report, court_opinion, news_index, election_calendar, policy_framework, party_policy, polling_data, regulatory_timeline, regulatory_consultation, international_standard, vendor_docs |
| update_cadence | Weekly |
## Sample JSON
```json
{
"report_id": "RPT_01KSHN11BNS8GSRWBRSNY2N409",
"slug": "ai-agents-legal-workflows",
"schema_version": "2026-05-28",
"title": "AI agents for legal workflows",
"abstract": "Legal AI agents should be bought as governed workflows: what they may read, draft, send, file, sign, remember, escalate and prove after lawyer review.",
"category": "AGENTS",
"sector": "Legal services",
"region": "GLOBAL",
"topics": [
"legal AI agents",
"legal AI providers",
"legal procurement",
"agent governance",
"legal retrieval",
"document intelligence",
"private evals",
"cross-border AI",
"AI regulation",
"legal AI litigation",
"AI politics"
],
"last_updated_at": "2026-05-31T00:00:00Z",
"source_count": 105,
"evidence_count": 118,
"confidence_score": 0.89,
"web_url": "/r/ai-agents-legal-workflows/RPT_01KSHN11BNS8GSRWBRSNY2N409",
"sample_endpoint": "/api/v1/reports/RPT_01KSHN11BNS8GSRWBRSNY2N409/sample",
"inventory_endpoint": "/api/v1/reports/RPT_01KSHN11BNS8GSRWBRSNY2N409/inventory",
"full_endpoint": "/api/v1/reports/RPT_01KSHN11BNS8GSRWBRSNY2N409/full",
"schema_endpoint": "/api/v1/reports/RPT_01KSHN11BNS8GSRWBRSNY2N409/schema",
"redaction": {
"status": "sample",
"paid_sections_redacted": true,
"preview_limits": [
"summary",
"table_of_contents",
"evidence_profile"
]
},
"coverage": {
"category": "AGENTS",
"sector": "Legal services",
"region": "GLOBAL",
"topics": [
"legal AI agents",
"legal AI providers",
"legal procurement",
"agent governance",
"legal retrieval",
"document intelligence",
"private evals",
"cross-border AI",
"AI regulation",
"legal AI litigation",
"AI politics"
],
"source_classes": [
"official_model_docs",
"privacy_terms",
"legal_platform_docs",
"industry_survey",
"agent_benchmark",
"professional_benchmark",
"retrieval_benchmark",
"legal_benchmark",
"legal_ai_analysis",
"bar_guidance",
"regulator",
"document_ai_docs",
"knowledge_platform_docs",
"legal_risk_database",
"job_posting",
"package_telemetry",
"issue_tracker",
"status_page",
"procurement_record",
"regulatory_registry",
"court_docket",
"archived_web",
"security_advisory",
"domain_registry",
"certificate_transparency",
"grant_database",
"civic_complaint",
"lobbying_disclosure",
"environmental_compliance",
"bid_protest",
"municipal_disclosure",
"workplace_safety",
"takedown_notice",
"bug_bounty",
"standards_body",
"expert_profile",
"evaluation_framework",
"case_study",
"demo_video",
"expert_commentary",
"industry_report",
"court_opinion",
"news_index",
"election_calendar",
"policy_framework",
"party_policy",
"polling_data",
"regulatory_timeline",
"regulatory_consultation",
"international_standard",
"vendor_docs"
]
},
"confidence": {
"overall_confidence": 0.89,
"source_quality": 0.9,
"source_recency": 0.92,
"source_density": 0.88,
"claim_verifiability": 0.88,
"geographic_coverage": 0.84,
"known_gaps": [
"No public benchmark fully captures end-to-end legal agent workflows under real matter permissions.",
"Vendor-reported legal AI workflow metrics remain unevenly reproducible from public data.",
"Cross-border legal AI benchmarks remain weaker than English-language US legal benchmarks.",
"Some high-alpha public sources remain fragmented by court, state, city, app store, board packet, insurer, or procurement system.",
"Polling, news-index, and party-policy sources are useful political-risk tripwires but weak evidence unless tied to primary legal or regulatory records."
]
},
"document": {
"object": "report_document",
"schemaVersion": "2026-05-28",
"title": "AI agents for legal workflows",
"subtitle": "The workflow is the product.",
"abstract": "Legal AI agents should be bought as governed workflows: what they may read, draft, send, file, sign, remember, escalate and prove after lawyer review.",
"category": "AGENTS",
"sector": "Legal services",
"region": "GLOBAL",
"topics": [
"legal AI agents",
"legal AI providers",
"legal procurement",
"agent governance",
"legal retrieval",
"document intelligence",
"private evals",
"cross-border AI",
"AI regulation",
"legal AI litigation",
"AI politics"
],
"sources": [
{
"id": "S1",
"label": "OpenAI - introducing GPT-5.5",
"sourceClass": "official_model_docs",
"accessedAt": "2026-05-31",
"status": "verified",
"access": "sample",
"transformChain": [],
"url": "https://openai.com/index/introducing-gpt-5-5/"
},
{
"id": "S2",
"label": "OpenAI - API data controls",
"sourceClass": "privacy_terms",
"accessedAt": "2026-05-31",
"status": "verified",
"access": "sample",
"transformChain": [],
"url": "https://openai.com/policies/api-data-usage-policies/"
},
{
"id": "S3",
"label": "Anthropic - introducing Claude Opus 4.8",
"sourceClass": "official_model_docs",
"accessedAt": "2026-05-31",
"status": "verified",
"access": "sample",
"transformChain": [],
"url": "https://www.anthropic.com/news/claude-opus-4-8"
},
{
"id": "S4",
"label": "Anthropic privacy centre - personal data in model training",
"sourceClass": "privacy_terms",
"accessedAt": "2026-05-31",
"status": "verified",
"access": "sample",
"transformChain": [],
"url": "https://privacy.claude.com/en/articles/7996885-how-do-you-use-personal-data-in-model-"
},
{
"id": "S5",
"label": "Google Cloud - Gemini on Vertex AI data governance",
"sourceClass": "privacy_terms",
"accessedAt": "2026-05-31",
"status": "verified",
"access": "sample",
"transformChain": [],
"url": "https://cloud.google.com/vertex-ai/generative-ai/docs/data-governance"
},
{
"id": "S6",
"label": "Microsoft - Azure OpenAI data, privacy, and security",
"sourceClass": "privacy_terms",
"accessedAt": "2026-05-31",
"status": "verified",
"access": "sample",
"transformChain": [],
"url": "https://learn.microsoft.com/en-us/legal/cognitive-services/openai/data-privacy"
},
{
"id": "S7",
"label": "Google Cloud - Gemini 3 Pro on Vertex AI",
"sourceClass": "official_model_docs",
"accessedAt": "2026-05-31",
"status": "verified",
"access": "sample",
"transformChain": [],
"url": "https://docs.cloud.google.com/vertex-ai/generative-ai/docs/models/gemini/3-pro"
},
{
"id": "S8",
"label": "Thomson Reuters - next-generation CoCounsel Legal beta",
"sourceClass": "legal_platform_docs",
"accessedAt": "2026-05-31",
"status": "verified",
"access": "sample",
"transformChain": [],
"url": "https://www.thomsonreuters.com/en-us/posts/innovation/cocounsel-legal-reimagined/"
},
{
"id": "S9",
"label": "Thomson Reuters - 2026 generative AI in professional services",
"sourceClass": "industry_survey",
"accessedAt": "2026-05-31",
"status": "verified",
"access": "sample",
"transformChain": [],
"url": "https://www.thomsonreuters.com/en/reports/2026-ai-in-professional-services-report"
},
{
"id": "S10",
"label": "LexisNexis - Lexis+ AI with Protégé",
"sourceClass": "legal_platform_docs",
"accessedAt": "2026-05-31",
"status": "verified",
"access": "sample",
"transformChain": [],
"url": "https://www.lexisnexis.com/en-us/products/lexis-plus-ai.page"
},
{
"id": "S11",
"label": "LexisNexis - Shepard's BriefCheck on Lexis+",
"sourceClass": "legal_platform_docs",
"accessedAt": "2026-05-31",
"status": "verified",
"access": "sample",
"transformChain": [],
"url": "https://supportcenter.lexisnexis.com/app/answers/answer_view/a_id/1102608"
},
{
"id": "S12",
"label": "Harvey - Legal Agent Benchmark",
"sourceClass": "agent_benchmark",
"accessedAt": "2026-05-31",
"status": "verified",
"access": "sample",
"transformChain": [],
"url": "https://www.harvey.ai/blog/introducing-the-legal-agent-benchmark"
},
{
"id": "S13",
"label": "Harvey - BLB Research",
"sourceClass": "agent_benchmark",
"accessedAt": "2026-05-31",
"status": "verified",
"access": "sample",
"transformChain": [],
"url": "https://www.harvey.ai/blog/blb-research"
},
{
"id": "S14",
"label": "Scale AI - PRBench",
"sourceClass": "professional_benchmark",
"accessedAt": "2026-05-31",
"status": "verified",
"access": "sample",
"transformChain": [],
"url": "https://openreview.net/forum?id=hZzGIqkl0F"
},
{
"id": "S16",
"label": "LegalBenchmarks.ai - legal AI evaluation framework",
"sourceClass": "legal_benchmark",
"accessedAt": "2026-05-31",
"status": "verified",
"access": "locked",
"transformChain": [],
"url": "https://www.legalbenchmarks.ai/legal-ai-evaluation-framework"
}
],
"evidence": [
{
"id": "E1",
"sourceId": "S8",
"claimIds": [
"C1"
],
"summary": "Thomson Reuters frames next-generation CoCounsel Legal around legal-source grounding, Westlaw and Practical Law access, and agentic workflows rather than generic chat alone.",
"status": "verified",
"confidence": 0.88,
"access": "sample"
},
{
"id": "E2",
"sourceId": "S10",
"claimIds": [
"C1"
],
"summary": "Lexis+ AI with Protégé emphasises LexisNexis legal sources, Shepard's citation treatment, uploaded document analysis, drafting, and workflow support.",
"status": "verified",
"confidence": 0.88,
"access": "sample"
},
{
"id": "E6",
"sourceId": "S12",
"claimIds": [],
"summary": "Harvey's Legal Agent Benchmark targets long-horizon legal agent work across more than a thousand tasks, practice areas, and expert-written rubrics.",
"status": "verified",
"confidence": 0.88,
"access": "sample"
},
{
"id": "E8",
"sourceId": "S14",
"claimIds": [],
"summary": "PRBench reports hard professional-reasoning gaps, including legal samples where transparency, uncertainty, and domain diligence remain weaknesses.",
"status": "verified",
"confidence": 0.88,
"access": "sample"
},
{
"id": "E10",
"sourceId": "S16",
"claimIds": [],
"summary": "LegalBenchmarks.ai's framework scores legal AI beyond accuracy, including strategic fit, functionality, robustness, security, data privacy, vendor risk, adoption support, and cost or resourcing.",
"status": "verified",
"confidence": 0.88,
"access": "sample"
}
],
"claims": [
{
"id": "C1",
"statement": "The best legal AI programme in 2026 is an agentic legal work stack: providers, models, sources, tools, permissions, evals, and human review selected by use case.",
"confidence": 0.9,
"evidenceIds": [
"E1",
"E2"
],
"access": "sample"
}
],
"colophon": {
"checkedAt": "2026-05-31",
"latestRevisionAt": "2026-05-31T00:00:00Z",
"reviewStandard": "expert_human_reviewed",
"kinetics": {
"speed": "fast",
"impact": 0.96,
"momentum": 0.91,
"businessImpact": "Legal AI agents are not ready to replace a junior lawyer end to end, but they can compress bounded legal work when sources, permissions, evals and approval gates are explicit.",
"rateOfChange": "Harvey LAB, LegalBenchmarks.ai task updates, AI Act dates, Thomson Reuters v ROSS and API data-control changes can reopen approved source, action and retention routes within a quarter.",
"from": "generalist models wrapped by legal SaaS",
"toward": "vertical legal AI routes with source, workflow, eval, audit and approval controls"
},
"watch": [
"23 June 2026: European Commission high-risk AI consultation closes; scope could change which legal AI uses need AI Act controls [E117].",
"2 August 2026: AI Act Commission enforcement powers for GPAI providers begin; agent routes may need stronger model documentation, copyright-policy or incident-reporting controls [E116].",
"Thomson Reuters v ROSS docket entries or appeal briefing; legal-RAG source-rights and indemnity diligence may change [E86].",
"Harvey Legal Agent Benchmark result release or LegalBenchmarks.ai task-leaderboard update adding wrong-tool, missed-source or approval-gate failures; private legal-agent eval cases may change [E112, E114].",
"OpenAI API data controls or Azure OpenAI data/privacy/security page change; retention, region, subprocessor or tool-action approval may shift [E17, E18]."
],
"gapAnalysis": "Independent end-to-end legal-agent benchmarks using real source licences, DMS/VDR/Office permissions, messy documents, cross-border rules, regulatory-change events and human approval gates."
},
"body": {
"sections": [
{
"id": "verdict",
"title": "Executive summary",
"access": "sample",
"dek": "The winning legal AI system is the one that knows when not to act.",
"claimIds": [
"C1"
],
"evidenceIds": [
"E1",
"E2",
"E6",
"E8",
"E10"
],
"blocks": [
{
"id": "verdict-narrative",
"evidenceIds": [
"E1",
"E2",
"E6",
"E8",
"E10"
],
"type": "section_narrative",
"lead": "Legal AI agents are not autonomous junior lawyers. They are bounded work systems that can search, draft, inspect spreadsheets, build risk registers, hand off tasks and prepare outputs for review. Once software can act, the risk is no longer only bad text.",
"body": [
"That is still a large change. A useful provider behaves sensibly when the law is stale, the source is missing, the matter is privileged, the spreadsheet has formulas, or the user asks it to send something.",
"Agentic legal AI changes the risk surface. A chatbot can be wrong. An agent can be wrong and also act. Source access, permission design, approval gates, audit logs and private evals sit at the centre of the approval decision.",
"Two rules now sit beneath the rest. Allocate responsibility across the value chain rather than let disclaimers stack onto the end user [E96]. Pick the level of human oversight from the action, not the vendor [E101]."
],
"detail": [
"Provider categories sell different assets. Legal-source owners sell authority, vertical platforms sell workflow, cloud providers sell control planes and model access, and internal teams sell fit to local matters.",
"The right selection process starts with the legal job, not the demo. Research, diligence, redlining, litigation review, legal ops and client delivery need different controls."
]
},
{
"id": "model-companion-link",
"evidenceIds": [],
"type": "report_link",
"reportId": "RPT_01KSHN11BMQVWKQ68GCHYNG1KY",
"slug": "ai-models-legal-work",
"label": "AI models for legal work",
"dek": "Use this alongside frontier, open-weight, Chinese, embedding, reranking, OCR and deployment model selection.",
"relation": "Related report"
},
{
"id": "file-processing-providers-report-link",
"evidenceIds": [],
"type": "report_link",
"reportId": "RPT_01KSHN11BPS6Z1Y922MK2184F7",
"slug": "file-processing-providers-legal-ai",
"label": "File processing providers for legal AI",
"dek": "Provider diligence for OCR, parsing, extraction, classification, table recovery, VDR ingestion, pricing, controls and parser evals.",
"relation": "Related report"
},
{
"id": "agent-action-approval-tree",
"title": "Agent action approval tree",
"caption": "Tool use turns answer risk into action risk.",
"evidenceIds": [
"E10"
],
"type": "decision_tree",
"lead": "Classify the action before choosing the agent. The risk changes when software can mutate the matter record.",
"steps": [
{
"prompt": "Is the agent only reading or comparing approved matter sources?",
"tone": "success",
"branches": [
{
"label": "Yes",
"outcome": "Allow with source logs, permission checks and no-answer behaviour.",
"tone": "success"
},
{
"label": "No",
"outcome": "Keep screening the action.",
"next": "Q2 drafting",
"tone": "monitor"
}
]
},
{
"prompt": "Will it draft, edit, extract fields or stage spreadsheet changes?",
"tone": "warning",
"branches": [
{
"label": "Yes",
"outcome": "Allow as a proposal with source links, diff view and reviewer assignment.",
"tone": "warning"
},
{
"label": "No",
"outcome": "Keep screening for external action.",
"next": "Q3 external action",
"tone": "monitor"
}
]
},
{
"prompt": "Can it send, file, sign, bill, tag privilege or change records?",
"tone": "blocked",
"branches": [
{
"label": "Yes",
"outcome": "Require explicit human approval before execution.",
"tone": "blocked"
},
{
"label": "No",
"outcome": "Run as a monitored workflow and sample failures.",
"tone": "success"
}
]
},
{
"prompt": "Will output become advice, privilege, citation or client position?",
"tone": "blocked",
"branches": [
{
"label": "Yes",
"outcome": "Require lawyer approval, source verification and retained reviewer notes.",
"tone": "blocked"
},
{
"label": "No",
"outcome": "Use sampling and regression evals to keep approval current.",
"tone": "monitor"
}
]
}
]
},
{
"id": "verdict-table",
"title": "The operating decision",
"evidenceIds": [
"E1",
"E2",
"E6",
"E8",
"E10"
],
"type": "table",
"columns": [
{
"key": "decision",
"label": "Decision",
"align": "left"
},
{
"key": "status",
"label": "Status",
"align": "left"
},
{
"key": "view",
"label": "View",
"align": "left"
}
],
"rows": [
{
"decision": "Buy versus build",
"status": "Route",
"view": "Buy mature workflows where source integration and adoption matter; build where proprietary data, client workflow, or sovereignty is the advantage."
},
{
"decision": "Source access",
"status": "Verify",
"view": "Prefer systems that rank authority, preserve currentness, and export evidence. A beautiful answer from the wrong source is still a liability."
},
{
"decision": "Agent autonomy",
"status": "Gate",
"view": "Allow read, draft, and propose. Gate send, file, sign, tag privilege, alter permissions, and mutate billing or spreadsheets."
},
{
"decision": "Benchmarks",
"status": "Test locally",
"view": "Use public benchmarks to design private evals; do not outsource procurement to vendor leaderboards."
},
{
"decision": "Unsupervised legal work",
"status": "Human gate",
"view": "No final advice, filings, privilege calls, or client-facing legal conclusions without lawyer approval and audit logs."
}
]
},
{
"id": "route-definition-callout",
"title": "Route definition",
"evidenceIds": [
"E1",
"E2",
"E6",
"E8",
"E10"
],
"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": "verdict-section-callout",
"title": "Approve actions, not demos",
"evidenceIds": [
"E1",
"E2",
"E6",
"E8",
"E10"
],
"type": "callout",
"tone": "info",
"body": "Any agent that can change, send, file, tag, bill or store work needs permission scope, logging and human approval before rollout."
}
]
},
{
"id": "market-map",
"title": "Agent problem space",
"access": "locked",
"dek": "Provider selection starts with the components of the work system, not a vendor list.",
"claimIds": [],
"evidenceIds": [],
"blocks": [
{
"id": "market-map-locked",
"title": "Locked section",
"evidenceIds": [],
"type": "callout",
"tone": "locked",
"body": "Purchase or subscribe to unlock this section."
}
]
},
{
"id": "provider-landscape",
"title": "Provider landscape",
"access": "locked",
"dek": "Choose providers by the work they can get through review.",
"claimIds": [],
"evidenceIds": [],
"blocks": [
{
"id": "provider-landscape-locked",
"title": "Locked section",
"evidenceIds": [],
"type": "callout",
"tone": "locked",
"body": "Purchase or subscribe to unlock this section."
}
]
},
{
"id": "source-access-retrieval",
"title": "Source access and retrieval",
"access": "locked",
"dek": "Retrieval is the chain of custody for legal reasoning.",
"claimIds": [],
"evidenceIds": [],
"blocks": [
{
"id": "source-access-retrieval-locked",
"title": "Locked section",
"evidenceIds": [],
"type": "callout",
"tone": "locked",
"body": "Purchase or subscribe to unlock this section."
}
]
},
{
"id": "agent-action-risk",
"title": "Agent action risk",
"access": "locked",
"dek": "Read and draft are safe defaults; send, file and sign need approval gates.",
"claimIds": [],
"evidenceIds": [],
"blocks": [
{
"id": "agent-action-risk-locked",
"title": "Locked section",
"evidenceIds": [],
"type": "callout",
"tone": "locked",
"body": "Purchase or subscribe to unlock this section."
}
]
},
{
"id": "private-evals",
"title": "Private evals",
"access": "locked",
"dek": "A private eval is the buyer's test set for real legal work.",
"claimIds": [],
"evidenceIds": [],
"blocks": [
{
"id": "private-evals-locked",
"title": "Locked section",
"evidenceIds": [],
"type": "callout",
"tone": "locked",
"body": "Purchase or subscribe to unlock this section."
}
]
},
{
"id": "roi-evaluation",
"title": "ROI / evaluation",
"access": "locked",
"dek": "Measure finished legal work, not token volume or seat count.",
"claimIds": [],
"evidenceIds": [],
"blocks": [
{
"id": "roi-evaluation-locked",
"title": "Locked section",
"evidenceIds": [],
"type": "callout",
"tone": "locked",
"body": "Purchase or subscribe to unlock this section."
}
]
},
{
"id": "workflow-playbooks",
"title": "Workflow playbooks",
"access": "locked",
"dek": "A legal agent should own one handoff, not the whole matter.",
"claimIds": [],
"evidenceIds": [],
"blocks": [
{
"id": "workflow-playbooks-locked",
"title": "Locked section",
"evidenceIds": [],
"type": "callout",
"tone": "locked",
"body": "Purchase or subscribe to unlock this section."
}
]
},
{
"id": "document-parsing",
"title": "Document parsing",
"access": "locked",
"dek": "Bad extraction turns a good agent into a confident guesser.",
"claimIds": [],
"evidenceIds": [],
"blocks": [
{
"id": "document-parsing-locked",
"title": "Locked section",
"evidenceIds": [],
"type": "callout",
"tone": "locked",
"body": "Purchase or subscribe to unlock this section."
}
]
},
{
"id": "privacy-deployment",
"title": "Contract risk in deployment",
"access": "locked",
"dek": "Training use, retention, region, audit, action logs and incident terms decide whether the demo is purchasable.",
"claimIds": [],
"evidenceIds": [],
"blocks": [
{
"id": "privacy-deployment-locked",
"title": "Locked section",
"evidenceIds": [],
"type": "callout",
"tone": "locked",
"body": "Purchase or subscribe to unlock this section."
}
]
},
{
"id": "cross-border",
"title": "Cross-border controls",
"access": "locked",
"dek": "The model may speak the language and still miss the law.",
"claimIds": [],
"evidenceIds": [],
"blocks": [
{
"id": "cross-border-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": "Use the pilot to decide what may scale.",
"claimIds": [],
"evidenceIds": [],
"blocks": [
{
"id": "implementation-plan-locked",
"title": "Locked section",
"evidenceIds": [],
"type": "callout",
"tone": "locked",
"body": "Purchase or subscribe to unlock this section."
}
]
},
{
"id": "operational-exhaust",
"title": "Operational diligence",
"access": "locked",
"dek": "The awkward public records often say more than the case study.",
"claimIds": [],
"evidenceIds": [],
"blocks": [
{
"id": "operational-exhaust-locked",
"title": "Locked section",
"evidenceIds": [],
"type": "callout",
"tone": "locked",
"body": "Purchase or subscribe to unlock this section."
}
]
},
{
"id": "legal-political-risk",
"title": "Legal-political risk",
"access": "locked",
"dek": "Court rulings, elections, policy posture, lobbying, and public sentiment can change the acceptable agent route.",
"claimIds": [],
"evidenceIds": [],
"blocks": [
{
"id": "legal-political-risk-locked",
"title": "Locked section",
"evidenceIds": [],
"type": "callout",
"tone": "locked",
"body": "Purchase or subscribe to unlock this section."
}
]
},
{
"id": "implementation-cases",
"title": "Implementation cases",
"access": "locked",
"dek": "Case studies are useful when they expose the operating model behind the result.",
"claimIds": [],
"evidenceIds": [],
"blocks": [
{
"id": "implementation-cases-locked",
"title": "Locked section",
"evidenceIds": [],
"type": "callout",
"tone": "locked",
"body": "Purchase or subscribe to unlock this section."
}
]
},
{
"id": "expert-watchlist",
"title": "Expert watchlist",
"access": "locked",
"dek": "Follow the people producing tests, tools, and failure analysis.",
"claimIds": [],
"evidenceIds": [],
"blocks": [
{
"id": "expert-watchlist-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": "Evidence gaps for procurement.",
"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": "Legal AI agents, platforms and document-infrastructure routes with public support for diligence.",
"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 sources and diligence anchors behind the 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": "SDK",
"meaning": "Software development kit: packaged code and tools developers use to connect a product to an API 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."
},
{
"term": "RAG",
"meaning": "Retrieval-augmented generation: search approved sources first, then pass selected evidence into the model."
},
{
"term": "OCR",
"meaning": "Optical character recognition: converting scanned or image text into machine-readable text."
},
{
"term": "VDR",
"meaning": "Virtual data room: a permissioned deal, diligence or transaction document repository."
}
]
}
]
}
]
}
},
"gap_analysis": "Independent end-to-end legal-agent benchmarks using real source licences, DMS/VDR/Office permissions, messy documents, cross-border rules, regulatory-change events and human approval gates.",
"kinetics": {
"speed": "fast",
"impact": 0.96,
"momentum": 0.91,
"businessImpact": "Legal AI agents are not ready to replace a junior lawyer end to end, but they can compress bounded legal work when sources, permissions, evals and approval gates are explicit.",
"rateOfChange": "Harvey LAB, LegalBenchmarks.ai task updates, AI Act dates, Thomson Reuters v ROSS and API data-control changes can reopen approved source, action and retention routes within a quarter.",
"from": "generalist models wrapped by legal SaaS",
"toward": "vertical legal AI routes with source, workflow, eval, audit and approval controls"
},
"watch_signals": [
"23 June 2026: European Commission high-risk AI consultation closes; scope could change which legal AI uses need AI Act controls [E117].",
"2 August 2026: AI Act Commission enforcement powers for GPAI providers begin; agent routes may need stronger model documentation, copyright-policy or incident-reporting controls [E116].",
"Thomson Reuters v ROSS docket entries or appeal briefing; legal-RAG source-rights and indemnity diligence may change [E86].",
"Harvey Legal Agent Benchmark result release or LegalBenchmarks.ai task-leaderboard update adding wrong-tool, missed-source or approval-gate failures; private legal-agent eval cases may change [E112, E114].",
"OpenAI API data controls or Azure OpenAI data/privacy/security page change; retention, region, subprocessor or tool-action approval may shift [E17, E18]."
],
"sample_findings": [
"The best legal AI agent stack combines source access, workflow fit, permissions, evals, audit logs, and lawyer approval."
]
}
```