Enterprise engagements built on real infrastructure. Platform functions designed and deployed. Each one with a clear problem, solution, and metrics story.
CFO-sponsored, multi-department AI program. Built on Glean and n8n. 4 departments, 5 agents in 6 months.
Generates jurisdiction-specific separation agreements across US, EMEA, and Australia. Validates proposed severance against policy before sign-off. The employment legal team's highest-priority automation target — roughly 100 agreements a year, each taking 2–4 hours manually.
Full detail page →Natural-language access to procurement and revenue-contract execution data, plus outside-counsel spend by firm and practice area. Replaced manual pulls across Tableau, spreadsheets, and spend-management tools. Leadership's first direct line of sight into legal operations data.
Full detail page →Automates repetitive Secretary of State filings across registered states. Frees the corporate team from recurring manual data entry across multiple state portals with different formats and deadlines.
Full detail page →AI-powered legal assistant embedded directly in the Cowork workspace. Answers legal and policy questions in context, connected to the company's knowledge base. Access controls enforce department-level data boundaries.
Full detail page →Designed and built as part of the Legal AI Operating System. Each follows the same governed pipeline: Input → Router → Evaluator → Scoring → Audit Trail.
Two-stage pipeline — router classifies incoming matters, evaluator scores across 5 dimensions with configurable rubrics. Programmatic scoring with full audit trail.
Metrics in Project Metrics →Multi-standard contract review against 30+ legal standards. Risk scoring, clause extraction, playbook alignment. Processes ~800 contracts/year across 3 clients.
Metrics in Project Metrics →Separation agreement generation, severance validation, and legal metrics access — the platform version of the enterprise patterns. Strict data isolation between employment and commercial legal.
Metrics in Project Metrics →Automated document review and risk flagging for M&A due diligence. Classifies documents, extracts key provisions, flags anomalies against expected patterns.
Metrics in Project Metrics →Tracks regulatory changes across jurisdictions, matches them against active matters and practice areas, surfaces relevant changes before they become compliance gaps.
Metrics in Project Metrics →Semantic search across the firm's precedent library and knowledge base. Surfaces relevant prior work, clauses, and matter context during new matter intake and contract review.
Metrics in Project Metrics →Quarterly AI impact reports with audit-trail backing. One-click export of time saved, cost avoided, and governance compliance per client. Turns RFP responses from anecdotes into data.
Metrics in Project Metrics →The tools the K&I Program Manager uses day to day. Portfolio visibility, ROI tracking, pipeline management, enablement, and client reporting — all built as part of the Legal AI OS operations layer.
Cross-client KPI cards: functions deployed, hours saved, cost avoided, net ROI, active users. Breakdown by function and practice group with trend sparklines. Speaks the partner's language — PEP, RPL, realization — not IT metrics. Directly maps to JD requirement: "Maintain a portfolio dashboard highlighting adoption, client outcomes, and return on innovation."
In Full Platform Reference →Cost impact engine (time saved × rate cards by practice group). Calibrated baselines with documented methodology, sample sizes, and confidence levels. Quality metrics tracked by override rate, not satisfaction surveys. Directly maps to JD requirement: "Develop frameworks to track AI solution performance, including time saved, cost impact, and quality metrics."
In Full Platform Reference →Kanban board tracking every AI initiative from Discovery → Build → Review → Graduated (or Cancelled). WIP limits, department visibility, feedback log per project. Directly maps to JD requirement: "Lead end-to-end delivery of AI-enabled client projects from concept through implementation."
In Full Platform Reference →Workshop deck, prompt engineering guide for lawyers, adoption playbook, AI literacy FAQ, client conversation pack, RFP response templates. Ships with every agent. Directly maps to JD requirement: "Partner with Learning & Development to upskill lawyers and staff in AI literacy."
In Full Platform Reference →Quarterly AI impact reports with audit-trail backing. One-click export per client. Governance packs for RFPs with ABA 512 compliance evidence. Technical FAQ for client CISOs. Directly maps to JD requirement: "Create client-facing materials and outcome reports demonstrating measurable AI value."
In Full Platform Reference →RAG-powered chat that knows the documentation, retrieves relevant context, and answers in plain language. 89 indexed help docs. Source references for verification. Offline fallback when connectivity fails. Built once, ported across platforms.
In Full Platform Reference →Standalone POC deployed as its own application. Configurable rubrics, staffing team matching, evaluation criteria that adapt to the firm. 24 automated tests. Proved the intake pattern before it was absorbed into Legal AI OS.
GitHub →Independent 4-dimension evaluation engine for Harvey-deployed agents. Safety, Bias, Accuracy, Compliance scoring with hard veto. Drift detection across 7 dimensions. Vendor can't grade its own homework.
In Full Platform Reference →Seven governance pillars — Confidentiality, Data Privacy, Traceability, Auditability, HITL, Model Governance, Explainability. Gated data flow pipeline. ABA Formal Opinion 512 compliance baked into evaluation prompts.
Governance Model page →