What did the ACC/Everlaw survey find about in-house legal AI adoption?

The ACC (Association of Corporate Counsel) and Everlaw GenAI Survey for 2026 found that corporate legal department AI adoption more than doubled in a single year: from 23% to 52%. More significantly for outside counsel, 64% of in-house legal teams now expect to depend less on outside law firms because of AI capabilities they are building internally. A parallel FTI General Counsel Report (March 11, 2026) found that 87% of GCs now use AI in their teams, up from 44% in 2025.

Which legal work is most at risk of being brought in-house because of AI?

The categories most exposed to in-house AI displacement are: standard contract drafting and review (NDAs, MSAs, vendor agreements), basic employment law advice (standard policy questions, routine agreement review), routine compliance work (GDPR/CCPA checklists, standard regulatory filings), legal research on well-established questions of law, and internal legal memos on recurring issues. These are high-volume, lower-judgment tasks that law firms have billed at full rates because they required a licensed attorney — and that AI now allows in-house teams to handle at dramatically lower cost.

What legal work is hardest to bring in-house even with AI?

The work most resistant to in-house displacement: complex litigation requiring deep court experience and local relationships, specialized regulatory counsel in niche industries (healthcare, financial services, energy, immigration), novel legal arguments without clear precedent, multi-jurisdiction matters requiring coordinated outside expertise, and high-stakes transactional work where outside counsel provides both legal advice and deal-making experience. Advisory work built on genuine specialization and long-standing relationships is the most defensible — not because clients can't do it themselves, but because the risk of error is high enough that they don't want to.

How should a small law firm respond to in-house AI advancement?

Three moves: (1) Audit which client relationships are based on delivering routine, high-volume work — those are the highest-risk relationships. Begin repositioning those relationships around advisory depth and specialized judgment. (2) Specialize explicitly. A firm with a narrow, deep practice area (estate planning for healthcare executives, employment law for staffing agencies, tax controversy) is harder to disintermediate than a generalist firm. (3) Get ahead of the conversation. The GCs who are building internal AI capabilities are sophisticated — they will appreciate a firm that proactively discusses how AI changes the engagement. Avoiding the topic is worse than addressing it directly.

Does this threat apply to accounting and consulting firms, or just law firms?

It applies to all professional services. Corporate finance teams with AI are now running their own variance analyses and preparing their own financial models — historically accounting firm deliverables. Consulting firms are facing clients who have built internal analytics capabilities that can run the same benchmarking analyses the consulting firm previously charged for. The ACC/Everlaw survey is specific to legal departments, but the mechanism is identical: clients who adopt AI internally become less dependent on the corresponding external professional service. Each firm type should audit which of their deliverables their clients can now do themselves.

In-House AI vs. Outside Counsel: ACC 2026

Published: March 14, 2026 | By: The Crossing Report | 7 min read

Summary

The ACC/Everlaw GenAI Survey for 2026 documents something small law firms — and their accounting and consulting counterparts — need to sit with: corporate legal AI adoption doubled in a single year, and 64% of in-house teams now expect to depend less on outside counsel because of AI they're building internally. This isn't a future threat. The clients building those capabilities are the same clients who hired you for the work AI is now replacing. Here's what's most at risk, what's defensible, and three moves worth making this quarter.

The Numbers

Corporate legal department AI adoption: 23% in 2025 → 52% in 2026. One year. More than doubled.

Of those departments using AI: 64% now expect to depend less on outside law firms because of AI capabilities they're building internally.

Source: ACC/Everlaw GenAI Survey, 2026.

A parallel data point from the FTI/Relativity General Counsel Report (March 11, 2026): 87% of GCs now use AI in their legal teams, up from 44% in 2025. Of those, 53% have formalized technology roadmaps, and approximately 70% plan to invest in new legal technology in the next 12 months.

These are the clients who currently give small and mid-size law firms their best work. The corporate GC who was previously dependent on outside counsel for standard contract review, routine employment advice, and basic legal research is increasingly doing those tasks in-house — not because she stopped valuing legal expertise, but because AI allowed her to scale her team's output without additional headcount.

What Is Actually Happening in These Legal Departments

The pattern the surveys document is not "clients replacing lawyers with AI." It's more specific: clients are using AI to absorb the high-volume, lower-judgment work they previously outsourced, while still relying on outside counsel for complex, specialized, and high-stakes matters.

The work moving in-house looks like:

Standard contracts: NDA drafting and review was one of the first legal tasks corporate departments learned to handle with AI. An in-house legal team with CoCounsel or Harvey can review a stack of NDAs for red flags and exceptions in a fraction of the time it would take to send them out and wait. The outside law firm's NDA work is gone.

Policy advice on recurring issues: "Does this termination comply with our state's employment law?" is a question that corporate counsel used to route to outside firms. With AI and a well-built internal knowledge base, in-house teams answer these questions themselves. The outside firm sees fewer employment advisory calls.

Routine compliance documentation: GDPR notices, standard vendor due diligence, routine regulatory checklists — all of this is now within reach of a well-equipped in-house team. The compliance work that previously flowed to outside firms is staying inside.

Legal research on established questions: When the issue is well-settled law (not a novel question), AI-assisted research closes the knowledge gap between a small in-house team and a large outside firm. The in-house team does the research and writes the memo.

The Work That Remains Defensible

The work that doesn't move in-house — the work that continues to flow to outside firms even as corporate departments build AI capabilities — shares certain characteristics:

Specialized expertise outside the GC's domain. A GC at a software company doesn't maintain deep expertise in healthcare regulatory compliance. When a software client expands into healthcare, they hire specialists. AI doesn't change this — it may make the GC smarter about the domain, but not expert enough to handle the work without experienced outside counsel.

High-stakes matters where the cost of error is existential. Bet-the-company litigation, major M&A transactions, regulatory investigations — the clients who use AI to handle routine contract review still want experienced outside counsel for the matters that could end the company. The judgment premium holds.

Local knowledge and relationships. A corporate client navigating a state court proceeding or a regulatory matter with a specific agency needs outside counsel who knows the judge, knows the regulator, knows the local practice. AI doesn't provide this. Network and relationships remain defensible.

Complex, multi-jurisdiction coordination. Large transactions or disputes spanning multiple jurisdictions require outside counsel who can coordinate across markets and relationships. In-house teams, even with AI, typically can't manage this coordination internally.

Novel legal questions without established precedent. When a client faces a genuinely new legal question — a regulatory gap, a new statute, an unsettled area of case law — AI is least helpful and outside counsel expertise is most valuable. The legal judgment premium concentrates at the frontier.

Three Moves for Small Law Firms This Quarter

Move 1: Audit Which Relationships Are Based on Routine Volume Work

Most small law firms have at least one or two clients where the relationship is primarily built around delivering high-volume, standardized work — the same contract types processed repeatedly, routine employment advice, recurring compliance documentation.

These are the relationships most at risk. The conversation with those clients is going to happen eventually. You can initiate it now, on your terms, or react to it when they stop sending the work.

Consider proactively asking: "We've been handling a lot of your [contract review / employment questions / compliance documentation]. We're seeing your industry change significantly with AI — how are you thinking about handling these workflows internally going forward? We'd like to make sure what we do for you stays aligned with where you need us most."

This is a good-client conversation, not a defensive one. It positions you as a trusted advisor who sees the same landscape they do.

Move 2: Specialize Explicitly

The firms most resistant to in-house AI displacement are the specialists. Not "employment law for companies" — but "employment law for staffing agencies," "tax controversy for real estate investors," "immigration law for technology companies."

Tight specialization makes you the outside counsel who knows things the GC's AI doesn't reliably know — industry-specific precedent, regulator relationships, deal-making experience in a specific transaction type, the judgment that comes from handling hundreds of similar situations.

If your firm does not have an explicit specialty that your ideal client could not easily describe to a colleague in one sentence, this is the quarter to define one.

Move 3: Get Ahead of the AI Conversation

The GCs building internal AI capabilities are sophisticated. They are watching which outside firms talk honestly about AI and which pretend the landscape hasn't changed.

A firm that proactively discusses its own AI use — how it uses Claude or CoCounsel or Spellbook internally, how that changes what you can deliver and at what speed — communicates competence to a GC who understands the technology.

A firm that avoids the AI topic or claims "we're exploring it" in 2026 signals that it hasn't adapted to the landscape the GC is already operating in.

The conversation is not: "AI is coming, let's talk about it." It's: "Here's how we've changed our practice to stay ahead of what your team can do internally. Here's what we can now do faster. Here's where we're still irreplaceable."

The Accounting and Consulting Parallel

The mechanism is the same across all professional services. The ACC/Everlaw survey measures legal departments, but the dynamic it documents is universal:

Clients who adopt AI internally become less dependent on the corresponding external professional service for the work AI can handle. They remain dependent on external professionals for the work AI can't handle.

For accounting firms: corporate finance teams with AI are doing their own variance analyses and financial models. The accounting firm's value is shifting from producing the analysis to interpreting it, from building the model to applying judgment to what the model shows.

For consulting firms: clients with internal analytics capabilities no longer need external consultants to run benchmark analyses. The consulting firm's value is shifting from delivering data to delivering the judgment about what to do with it.

The question every professional services firm should be asking right now: which of our services could our clients' AI now handle without us? And what do we offer that they can't replicate?

The answers to those two questions are your firm's strategic priorities for the next 24 months.

64% of Your Biggest Clients Are Building AI to Replace the Work You Do — Here's What Survives