Our 2026 survey of legal technology leaders — spanning boutique, midsize and AmLaw firms across North America — captures where AI, automation and IT service management are actually landing in practice, not where marketing decks claim they are. The headline is clear: adoption is accelerating on almost every dimension, but the gap between firms with a coherent plan and firms without one is widening at a pace that has begun to alarm the partners we brief.
AI has moved from experiment to expectation
More than three-quarters of respondents now report at least one AI capability in production, up sharply year over year. Copilot and Gemini deployments lead by volume, followed by matter-specific agents in discovery, drafting, intake and matter budgeting. Clients are asking about AI directly in RFPs, and an increasing share of outside-counsel guidelines now explicitly require it — with governance controls attached. The window in which 'we're evaluating AI' was an acceptable answer to a sophisticated GC has effectively closed.
Automation is quietly reshaping the back office
Beyond generative AI, firms are automating timekeeping narratives, conflicts, billing review, vendor management, matter opening and closing procedures, and executive reporting. The firms furthest along report double-digit gains in realization and measurable reductions in write-offs — quiet, unglamorous wins that add up to real margin over a fiscal year. This is where the fastest payback in legal tech currently lives, and it does not require a firmwide AI strategy to start capturing it.
Service management is the unlock
The survey's strongest correlation, by a wide margin, is between mature IT service management and AI ROI. Firms with clear SLAs, disciplined change management, a single service desk, real problem management and monthly executive reporting get meaningfully more from every AI dollar. Firms without that operational spine stall — regardless of how much they spend on models, licenses or consultants. If we had to pick a single lever to move first, this would be it.
Cybersecurity has become inseparable from AI
Respondents increasingly report that AI decisions are cybersecurity decisions. Every new capability expands the tenant surface area, every new agent introduces new privilege questions, every new integration is a new logging surface. Firms that had already invested in modern identity, EDR and 24×7 detection are able to green-light AI faster because their risk committees can see how the new capability lands inside the existing control framework. Firms that hadn't invested in that foundation are discovering that AI is now blocked on it.
The 2027 gap will be uncomfortable
Firms that treat AI, automation, cybersecurity and service management as one connected program — with a single governance body, a single reporting cadence and a single accountable executive — are pulling ahead on margin, talent retention and client wins. Firms that treat them as four separate initiatives are already falling behind, and the gap compounds each quarter. Partners we brief consistently underestimate the compounding, and consistently regret that they did.
What to do with these findings
Use the report as a benchmark, not a scorecard. Read the sections that describe firms one or two tiers ahead of yours, identify the two or three practices that would produce the highest payoff at your firm, and stand up an execution plan against them by next quarter. The firms that will benefit most from this survey are not the ones that read it — they are the ones that act on it within ninety days.
See the benchmarks above for benchmarks by firm size and practice area, or talk to our team about where your firm sits on the curve. Subscribe below to get next quarter's data as soon as it drops.
Contact us
