AI and Outside Counsel: The Gap Between What GCs Expect and What Firms Deliver

General Counsel overwhelmingly expect AI to lower what they pay outside counsel, but few have seen it on a bill or written it into their outside counsel guidelines, and fewer than half have raised AI with a firm.

Illustration of the gap between what general counsel expect from AI and what outside law firms deliver
Key findings from the AdvanceLaw GC survey
  1. Nearly nine in ten GCs expect AI-driven efficiency to show up in outside counsel bills; almost none have seen it.
  2. Only a small handful of legal departments have added explicit AI language to their outside counsel guidelines.
  3. Fewer than half of GCs have raised AI explicitly with a firm.
  4. On whether AI will be a real factor in choosing or managing outside counsel within the next year, about a third say yes, a third say no or not yet, and the rest say it depends on the matter.

General Counsel are nearly unanimous that artificial intelligence should be changing their relationships with outside counsel. What AdvanceLaw hears across conversations with legal departments throughout our network tells a more complicated story: expectation has outpaced action almost everywhere it’s been tested. Billing hasn’t moved. Outside counsel guidelines haven’t been rewritten. And the conversation about AI, as explicit as it would need to be, mostly isn’t happening yet. Here’s where things actually stand.

The Billing Evidence Gap: Are GCs Seeing AI Savings on Outside Counsel Bills?

Nearly nine in ten GCs expect AI-driven efficiency to eventually show up in their outside counsel bills, whether as lower rates, faster turnaround, or tighter scope. Almost none have actually seen it. Only a small fraction can point to a specific, identifiable AI line item on an invoice, and even then the savings typically fall short of what was promised.

The expectation isn’t unreasonable. Firms genuinely are using AI tools internally. What’s missing is the translation from efficiency gained to value passed through. Until that shows up in writing, on a bill, “using AI” and “billing differently because of it” are two separate claims, and most legal departments have only heard the first one.

Outside Counsel Guidelines Are Playing Catch-Up on AI

Only a small handful of legal departments have added explicit AI language to their outside counsel guidelines or engagement terms. Some say they are actively rewriting theirs right now. The large majority of legal departments haven’t made any updates.

That gap matters most around confidentiality and data governance. Most departments still lean on existing confidentiality and professional responsibility obligations rather than AI-specific contract language, but this is the area GCs flag most often as the next thing to get addressed. The focus will be on which tools are in use, on which matters; whether client data trains third-party models; whether work happens in closed, non-training environments.

The Expectation Is Universal. The Conversation Is Not.

Here’s the disconnect in two numbers: the overwhelming majority of GCs expect AI savings to show up in their bills, but fewer than half have raised AI explicitly with a firm. The expectation exists quietly, assumed rather than asked for, which is part of why it hasn’t been met.

Asked whether AI will be a real factor in choosing or managing outside counsel within the next year, the room splits roughly three ways: about a third say yes, unequivocally; a third say no, or not yet; the rest say it depends entirely on the matter. Nobody’s betting against AI mattering. They’re just not agreed on when, or where, it matters first.

Where This Goes Next

The appetite for AI is concentrated where the stakes and the document volume are both high.

Bar chart of where GCs most want AI applied: Litigation & E-Discovery 63%, M&A & Due Diligence 47%, Contract Drafting & Review 26%, IP & Patent 16%, Regulatory Tracking 5%.
Source: AdvanceLaw GC survey. Litigation & E-Discovery leads at 63% of GCs, followed by M&A & Due Diligence at 47%. Contract drafting and review (26%), IP and patent (16%), and regulatory tracking (5%) trail well behind.

 

But wanting AI in litigation and M&A is different from firms actually deploying it there. The tools GCs report their firms using skew general-purpose, not practice-specific.

Bar chart of AI tools GCs are using: Microsoft Copilot 58%, Claude 42%, ChatGPT Enterprise 42%, Harvey 37%, CoCounsel 32%, Legora 16%.
Source: AdvanceLaw GC survey. Microsoft Copilot leads at 58%, followed by Claude and ChatGPT Enterprise at 42% each. Practice-specific tools trail: Harvey (37%), CoCounsel (32%), and Legora (16%).

 

Put the two charts together and a pattern emerges: most firms have adopted a general-purpose assistant and stopped. Fewer have built the practice specific workflows, document review pipelines, or litigation-support tooling that would actually move the needle in the areas GCs care most about. That’s optimization, not innovation, and it’s the gap GCs are asking their firms to close.

What Legal Should Do First for AI in relation to Outside Counsel

Three things worth doing this quarter, regardless of firm size or practice mix:

  1. Put disclosure and data-use language in your outside counsel guidelines. Require firms to say which AI tools they use and how client data moves through them. This is the fastest-moving gap, and the one GCs most often say is coming next.

  2. Make “show me the savings” a standing question, not a one-time one. Ask firms at the beginning of a new matter how AI will be factored into staffing, scope, or price, every time, not just when a firm brings it up first.

  3. Ask for a specific example, not a yes-or-no. “Do you use AI?” invites a marketing answer. “Show me one matter where AI changed your staffing, timeline, or price” invites an honest one.

Frequently Asked Questions

Do outside counsel guidelines need AI-specific language?

Most legal departments haven’t added it yet. Only a small handful have, and GCs most often flag confidentiality and data governance as the next area to address: which AI tools are used, on which matters, and whether client data trains third-party models.

Are law firms passing AI savings on to clients?

Not yet. Nearly nine in ten GCs expect savings to show up in their bills, but only a small fraction can point to an identifiable AI line item, and the savings typically fall short of what was promised.

What should a GC ask a law firm about AI?

Ask which AI tools the firm uses and how client data moves through them, and ask for one example of a matter where AI changed staffing, timeline, or price.

Will AI influence which outside counsel GCs hire?

GCs are split on timing: about a third say yes, a third say no or not yet, and the rest say it depends on the matter.

Want to see how AI can support your legal department? Explore Mitratech’s AI solutions for legal operations or get the AI in legal eBook.