Use of AI in law has grown rapidly (74% of legal professionals now use AI multiple times a week, yet only 35% say an active strategy is visible in daily work), but there is significant variance in adoption maturity across law firms based on a variety of factors including firm size, practice areas, workflow, technical expertise, and systems readiness.

Law firms are also using AI in more sophisticated ways – from beginners using off-the-shelf general-purpose models and tools, to the growing adoption of legal AI platforms like Harvey and Legora, to more advanced firms leveraging in-house technical teams to build custom AI harnesses, fine-tune AI models and procure specialized infrastructure. Maturity depends on how tools are integrated, developed, and governed, regardless of whether they are bought or built. 

Tracking these shifts by practice area and workflow is critical to establish best practices and quantify outcomes, ROI, and cost profiles creating a definitive benchmark for AI maturity and value in law. This work on creating a baseline is particularly important now because AI usage in all sectors including Law is expected to increase dramatically as both human-directed chatbot use and automated agentic workflows expand. 

Goldman Sachs research projects global monthly token consumption to increase from roughly 5 quadrillion tokens in 2026 to 120 quadrillion tokens by 2030, a 24-fold increase driven mostly by agentic workflows. These global cross-industry indicators suggest that AI activity, cost, and complexity will scale faster than seat counts, making it urgent for law firms and legal departments to understand usage, cost, complexity, ROI, and outcomes at the workflow level.

Phase 1: Interviews & Study Scoping

• Present study objectives, participation benefits and technical guardrails to cohort of CodeX approved law firms. Confirm 3-5 participant firms.
• Conduct survey with IT teams of participating firms to confirm AI usage – detailing providers, use cases and available metrics
• Publish measurement toolkit, AI taxonomy, maturity framework
• Select practice areas and priority workflows per firm.

Phase 2: Installation & Measurement Setup

• Private, secure installation of measurement toolkit as needed by each participating firm
• Refine CodeX categorization of AI Maturity, Legal Processes and Workflow taxonomy to ensure accurate internal benchmarking within and across firms
• Classify historical data and forward data collection on a continuous basis

Phase 3: Analysis & Recommendations

• Analyze historical data, identify initial trends by practice and workflows within each firm for internal benchmarking, provide report to each firm 
• Determine firm specific ROI metrics where reliable data is available
• Develop recommendations for cost optimization by firm

Aggregate, anonymized results will be shared with CodeX for research and teaching based on permissions granted by participating firms.