By The LexText Team
Updated:
Jul 19, 2026
Four GCs from leading companies asked us to show them what's behind the curtain of AI in litigation. What they wanted to know had less to do with cost savings than everyone assumes. Here are the five questions every GC should be asking outside counsel.
The GC mindset: litigation is a problem to manage, not a project to fund
Our CEO spent a decade in the general counsel seat, so this was familiar territory.
When a complaint lands on a GC's desk, it is not an intellectual exercise, and it is certainly not a fee opportunity. It is a problem to manage. The job is to quickly assess the merits, resolve the matter as efficiently as the facts allow, and make one critical judgment early: is this noise, or is this something the board needs to hear about? Then comes the next decision, which firm is the right one to handle it.
Every dollar and every week spent on litigation is a dollar and a week not spent on the business. That framing shapes everything a GC thinks about legal AI. Which brings us to the question every GC asks first.
The obvious question: are the savings reaching me?
The GCs in our session knew, at least in outline, what AI can now do in litigation. Draft answers to complaints. Prepare discovery requests and responses. Build the first version of a motion to dismiss. Analyze a complaint against the controlling standards and flag the arguments worth making. Work that once consumed dozens of associate hours now takes a fraction of the time when the tooling is built for litigation rather than adapted to it.
So the first question is exactly the one you would expect: if our outside counsel is using these tools, where are the savings going?
It's a fair question, and firms should expect to hear it in every rate negotiation and every RFP from now on. Clients have read the same efficiency headlines the firms have. They know the math is changing.
But here is the part that surprised us. Cost was not actually the headline of the conversation. Quality was.
The real headline: quality, and what a billable hour should buy
None of the GCs in that room objected to premium rates. They hire elite outside counsel precisely because elite judgment is worth paying for. Their frustration was never the rate. It was paying that rate for work that does not require it.
They do not want to pay $900 an hour for document review and weak first drafts. They do not want to be the ones funding the on-the-job training of first-year associates. They want to pay $900 an hour for judgment. Strategy. Seeing around corners. The instinct, built over hundreds of matters, for which argument wins and which settlement window is real. The skills that justify the rate.
This is what AI actually changes when it is built for litigation. The routine, lower-value work gets done faster and more accurately. Associates skip past the drudgery and start doing higher-level lawyering earlier in their careers. Partners spend their hours on the work that actually moves cases. And clients get better outcomes, not just cheaper invoices.
One of the strongest signals we've seen came from a litigation associate at a leading firm after weeks of daily use: the tools "level up my work." Same hours, higher-level lawyering. That is the trade GCs want to make, and it is a trade firms should want to make too.
Five questions every GC should ask outside counsel about AI
The GCs we met were preparing to bring this conversation to their firms directly. If you're in-house, these are the questions worth asking. If you're at a firm, these are the questions worth having answers to before the client asks.
What AI tools are you using on our matters, and what were they built for? General-purpose chatbots and litigation-native platforms are different animals. Ask specifically what the tool was designed to do and how it fits the litigation workflow.
How do you verify AI output before it reaches us or a court? Citation accuracy is the threshold issue. Ask what systematic verification exists, not just "a lawyer reviews everything." Every sanctions headline involving a hallucinated case involved a lawyer who was supposed to review everything.
How does your use of AI show up in our bills? If drafting an answer takes a third of the time it took two years ago, the invoice should reflect that somewhere: lower hours, alternative fee structures, or partner-level attention that was not economical before.
What is your data security posture for AI tools? SOC 2 certification, zero-data training policies, and clear data processing agreements should be table stakes for any tool that touches client information.
How are you using the time AI frees up? This is the question that separates firms that treat AI as a cost story from those that treat it as a quality story. The best answer sounds like: more time on strategy, earlier case assessment, deeper motion practice, faster responsiveness.
Notice what these questions do. They move the conversation past "do you use AI" (everyone will soon say yes) to "does your use of AI make you a better firm for us" (very few can answer well today).
What this means for law firms
The firms that win the next decade of work will not be the ones that resist this shift, nor will they be the ones that quietly pocket the efficiency gains. They will be the ones that walk into the pitch already ahead of the client's questions, with a clear account of what tools they use, how they verify output, and how the client benefits in both cost and quality.
Innovation is no longer a marketing slide. It is becoming a client requirement, written into outside counsel guidelines and RFP scoring criteria. The GCs we met were not hostile to their firms. They were excited to have this conversation with them. The opportunity for firms is to be ready for it.
What this means for GCs
The curtain is coming back, whether firms are ready or not. GCs who understand the current capabilities of litigation AI, what it does well, where it still needs expert judgment, and what verification looks like will negotiate better engagements, spot weak answers, and get more value from the counsel they already trust.
That is exactly why we built LexText the way we did: litigation-native, grounded in the actual workflow, with citation verification built in rather than bolted on. It is also why we spend as much time talking with in-house teams as we do with firms. Both sides of the relationship are asking the same question now, and the answer is better when they ask it together.
Want to stay ahead of what’s actually happening in AI and litigation? Follow LexText AI on LinkedIn, where we share new capabilities, court rulings on AI use, and the questions in-house teams and firms are asking each other. Our weekly briefing, Behind the Curtain, is coming soon.
Curious what litigation-native AI looks like in practice? Book a demo.