Annual DISCO Legal AI survey finds legal AI adoption up sharply on every measure — and an industry that spent last year forecasting savings now unable to forecast costs

DISCO (NYSE:LAW), a creator of industry-leading litigation technology, today announced the findings of its annual study of legal AI trends in law firms and corporate legal departments, conducted by Ari Kaplan Advisors. The research finds that legal teams have largely settled the question of AI’s capabilities and have moved on to a harder one: what it costs, and whether anyone can predict that number in advance.

In DISCO’s 2025 study, cost savings were a headline expectation, with 32% of participants reporting they were already realizing savings from AI and 55% expecting them. In the 2026 research, participants raised a different subject without being asked: the shift in legal AI pricing away from annual subscriptions toward consumption-based token and credit models, and the difficulty of budgeting against it. "The token/credit costs seem to be changing as the definition of a task evolves; there seems to be a lack of visibility into what constitutes a task and how credit usage is calculated," said one law firm participant.

Legal departments described building operational machinery around the problem in an effort to increase cost certainty. One reported billing token costs back to individual business units. Another is assigning each lawyer a personal token allocation, and was candid about the risk, noting that "this could result in much higher long-term costs since most people do not monitor their usage." A law firm participant framed it as a planning problem: "budgeting will become an even bigger challenge because it will be harder to identify what a tool should be used for, how much it costs, and the best way to budget for it."

"The Legal industry has resolved the AI capability question much faster than it has resolved the economics question. Legal teams have a greater understanding of what AI can do but very little visibility into what it will cost them to do it. Pricing transparency has become a bigger constraint on adoption than output, and both law firms and corporations are increasingly demanding better line of sight to both return on investment and the bottom line," said DISCO’s Director of AI Consulting, James Park.

Download the full survey results and white paper here.

Other important trends the survey identified include:

Clients are moving faster than their outside counsel

The pressure to adopt is increasingly coming from the top of the client organization. Pressure from the board or C-suite rose from 18% to 47% of corporate legal respondents in a single year, the largest single-year shift in the research, while pressure from the highest levels of company leadership rose from 64% to 76%. Law firms feel it downstream, with the share adopting legal AI specifically to meet rising client demand rising from 39% to 64%.

Several in-house participants described the relationship shifting beneath their outside counsel. "We are seeing the role of outside counsel changing from an originator of work to a validator," said one in-house lawyer.

Agentic AI has arrived — but almost no one thinks it is the endgame

Agents have moved from the margins into production. Sixty-two percent of participants report using AI agents for work including horizon scanning across jurisdictions, weekly legal hold reporting, data extraction, fact chronologies and document review.

Only 10%, however, believe legal AI is becoming exclusively agentic, and the reservation is about control rather than capability. "I hope legal use does not become exclusively agentic AI because that would mean we would lose control," said one law firm leader.

AI Review has passed its tipping point

Confidence in AI for document review has passed a threshold. Seventy-two percent of respondents rate themselves as somewhat or very confident using legal AI for document review compared with conventional manual review, up from 53% a year ago. Fifty-eight percent have integrated legal AI into routine legal processes, up from 35%, and 94% work in organizations with a formal AI policy, up from 80%. "I test it regularly and have verified it against managed review onshore and offshore; it has worked so well that I would never go back to a manual review for first- or second-pass review," said one participant.

The research also found discovery workloads rising for 53% of respondents, up from 47%, while the share reporting that complex litigation is taking longer fell from 39% to 34%. Several participants attributed rising volume to AI lowering the barrier to filing rather than to AI slowing matters down. "Due to AI, the bar for entering litigation from a demand letter to pleadings is much lower because anyone can engage in that process with AI’s help," said one respondent.

"The market is moving so quickly, and I hope this research helps legal teams in law firms and law departments benchmark their performance to drive their organizations forward," said Ari Kaplan, a legal industry analyst and the Principal of Ari Kaplan Advisors. "Leaders are now balancing the promise of legal AI with the practical realities of embedding it into the fabric of their operations."

Additional findings from the 2026 Legal AI Survey:

Pressure to adopt is now top-down on both sides. Seventy-six percent of corporate legal respondents report pressure from the highest levels of leadership, 55% from within the legal department and 27% from business units; 47% cite the board or C-suite. Sixty percent of law firm respondents report pressure from leadership and clients, 38% from partners and 26% from associates.
Efficiency dominates the rationale. Ninety-one percent of firm respondents and 92% of in-house respondents name efficiency as the primary driver. Sixty-four percent of firms are responding to rising client demand.
Discovery volume is rising while case duration is not. Fifty-three percent report an increased discovery workload and 31% report it unchanged; one respondent reported a decrease. Thirty-four percent say complex litigation is taking longer, and of those, 63% are investing in legal AI to offset it. Several participants attributed rising volume to AI lowering the barrier to filing.
In-house teams are reclaiming work. Participants described using agents and generative AI to keep matters in-house, including one who reported saving roughly $10,000 in outside counsel fees per M&A transaction by generating board resolutions from an internal archive.
Adoption is uneven. As one legal operations leader put it: "It's a journey, and we are at the end of the beginning." A law firm leader offered that legal AI is "not ubiquitous like email yet, though it is headed in that direction."

Methodology

In the summer of 2026, Ari Kaplan Advisors conducted an online survey completed by 104 individuals, supplemented by 33 in-depth interviews. Fifty-three respondents (51%) work in law firms and 51 (49%) in corporate legal departments. Sixty-eight percent work in organizations with more than 1,000 employees, and 91% of law firm participants handle commercial litigation, among other practice areas. Findings reported separately for law firms or corporate legal departments are based on approximately 51 to 53 responses. The prior-year study, conducted from August 8 through September 29, 2025, surveyed 112 individuals and interviewed 32 leaders, divided evenly between law firms and corporate legal departments.

About DISCO

DISCO (NYSE:LAW) provides comprehensive, innovative solutions for modern litigation. We create and service an intuitive, cloud-native platform at the forefront of litigation technology, backed by the partnership of expert professional services and support. Leveraging the latest in AI to help law firms and corporations achieve smarter outcomes faster, our scalable products and tools allow customers to simplify everyday tasks and tackle complex matters at every stage of litigation. Learn more at www.csdisco.com.

View source version on businesswire.com: https://www.businesswire.com/news/home/20260923427226/en/

Press: media@csdisco.com