In October 2025, the Association of Corporate Counsel and Everlaw reported on a survey of 657 in-house legal professionals in 30 countries. Among U.S. respondents, 52 percent were actively using generative AI, up from 23 percent a year earlier. The same survey found that 59 percent had seen no generative AI savings from their law firms yet.
Whether in-house teams are ahead of law firms depends on which survey you read. The Thomson Reuters Institute's 2025 Generative AI in Professional Services report, fielded in early 2025, put generative AI use at 28 percent for law firms and 23 percent for corporate legal departments. The two surveys use different samples and definitions, so neither settles who is faster.
What is clearer is the incentive. Law firms that bill by the hour earn less on a task when technology makes that task faster, unless they change how they price. In-house legal teams are judged on legal spend, on the speed of support to the business, and on the quality of legal outcomes, not on hours worked. Technology that reduces attorney time per matter is good for an in-house team on every one of those measures.
That asymmetry shapes what each side adopts, how fast, and what the in-house/outside counsel relationship looks like next.
The Incentive Asymmetry
Legal AI tools reduce the time required to complete legal work. For in-house teams, less time per task means more capacity, more work delivered at the same cost, and, when that capacity displaces outside counsel spend, a direct financial return. For law firms billing by the hour, less time per task means less revenue per task if the billing rate stays constant.
That creates an asymmetry in adoption incentives:
In-house legal teams have direct reasons to adopt AI tools that reduce attorney time: lower operating costs, faster support to the business, and more legal work handled with the same headcount. Every attorney hour saved can be redirected to higher-value work or can reduce outside counsel spend. Those are results a CFO can see in the budget.
Law firms have mixed incentives. The long-term competitive argument for adoption is real but diffuse. The short-term economics are harder: tools that let junior associates finish work faster reduce the billable hours those associates generate. Where partner compensation is tied to revenue and utilization rates matter for leverage, the billing model blunts the incentive to adopt efficiency-improving technology.
Managing partners are not irrational. They are responding to the incentives their business model creates. Firms that move to fixed fees or other alternative pricing remove much of that conflict, which is one reason to ask your firms how they price AI-assisted work.
Where In-House Teams Are Applying AI
In-house teams are applying AI in five areas.
Contract analysis and management. AI-assisted contract review, extraction, and risk analysis is the most common starting point in in-house legal. The financial case rests on one number: outside counsel spend on standard contract review before and after deployment. Track it from the first month, so the return is measured rather than estimated.
Legal research. AI research tools that synthesize case law, regulatory guidance, and secondary sources around a specific question can shorten work that used to require an outside counsel memo. Measure turnaround time on research questions before and after, and check the outputs: these tools can produce citations that do not exist. In Mata v. Avianca (S.D.N.Y., June 2023), a federal judge sanctioned two lawyers and their firm for filing a brief built on fabricated cases a chatbot had produced. Every authority still needs to be verified by a lawyer.
Compliance monitoring and regulatory tracking. AI systems that monitor regulatory developments across jurisdictions, flagging changes in rules, agency guidance, and enforcement trends that matter to the company's operations, can reduce the monitoring load that otherwise needs attorney time or an outside subscription.
Outside counsel management. Analytics on outside counsel invoices, such as billing anomalies, rate comparisons across matters, and matter budgets against actuals, give the in-house team better data when it negotiates with firms.
Matter management and operations. Tools that automate matter intake, standardize legal hold procedures, organize document management, and support legal project management reduce administrative work and make the department's own workload measurable.
Why Law Firm Change Is Slower
Law firm adoption is not zero, and on some measures it is ahead of in-house. Even so, firms face constraints beyond the billing model that in-house teams do not.
Partnership governance. Major decisions in a partnership can need consensus among dozens or hundreds of equity partners. Reaching agreement on AI strategy, vendor selection, and implementation takes time.
Professional responsibility. Law firms are responsible for the legal work they deliver to clients. Questions about whether AI-assisted work meets the standard of care, and who bears liability if it does not, are a real governance problem. An in-house lawyer using AI for internal work faces the same duties but a different risk profile.
Client pricing pressure. In the ACC/Everlaw survey, 50 percent of in-house respondents expected generative AI to lower their outside counsel costs. Clients that expect lower fees when AI shortens a task make the revenue effect of adoption uncertain for firm leadership.
Talent model dependencies. Large-firm economics depend on leverage: partners supervising and billing the work of junior associates. Tools that remove junior associate hours also remove part of the training ground for the associate-to-partner pipeline, which raises questions about talent development as well as revenue.
None of these concerns is irrational. Together they give in-house teams fewer obstacles to adopting AI than the firms that serve them.
What It Means for the In-House/Outside Counsel Relationship
In-house teams expect to send out less work. In the ACC/Everlaw survey, 64 percent of respondents expected to rely less on outside counsel because of generative AI. Routine work that was sent out for lack of capacity is the first candidate to stay in-house. High-complexity, high-stakes work is a different matter.
In-house teams are better placed to challenge outside counsel work. A general counsel whose team uses AI research tools has a better basis for judging the quality and efficiency of an outside research memo, and for questioning invoices that bill hours an AI-assisted team would not need. The information gap that favored firms in billing discussions narrows.
Firms will compete more on expertise than on effort. The work AI does least well is the work firms are best placed to sell: deep specialization, strategic judgment, institutional relationships, trial experience, and regulatory expertise beyond what AI-assisted research produces. Expect firms to use AI on the analytical work and price the judgment separately.
Some firms may build capabilities in-house teams cannot. A firm that sees patterns across many clients and industries could, with client consent and appropriate safeguards, turn that view into insight no single in-house team can see. Whether firms can do this within their confidentiality obligations is a question for their own professional responsibility advisers.
What General Counsels Should Do Now
For in-house legal leaders, the incentive advantage only pays off with deliberate investment.
Build an AI legal technology roadmap. Rather than adopting tools as vendors pitch them, rank use cases by financial impact, implementation complexity, and strategic importance. Contract analysis has the most directly measurable return, because it maps to outside counsel spend you already track.
Measure the return explicitly. Hold legal AI to the same financial discipline as other operational technology: a projected return, an implementation cost estimate, and tracking against the projection after go-live. State in advance what result would make you stop. Stopping is a legitimate outcome of a measured pilot.
Put AI use into outside counsel relationships. Address it in engagement letters and fee arrangements. Ask each firm whether it uses AI on your matters, on what tasks, and how the time saved shows up in the bill. Ask firms that do not use AI to explain the longer timelines and higher costs that follow.
Build legal technology capability in the department. A legal operations function with real expertise in AI deployment and data analysis, beyond contract-system administration, will make better buying decisions than one that treats technology as an IT responsibility.
Start with one number: last year's outside counsel spend on standard contract review, by matter type. If you cannot produce it, building that baseline is your first project, before any AI purchase.
GradeCircle builds and runs six products of its own: AfroSynergy, CallerBridge, HealthSquire, Heyzl, Lexfora, and LukeAPI, two of them in beta. The assumptions in our models come from operating those systems, not from a framework. See what we run.
If you have a specific decision in front of you, scope it in five questions and see the engagement, price band, and timeline before sharing contact details.
