How In-House Legal Should Negotiate Law Firm Fees in the AI Era

"You are using AI, so it has to be cheaper."

That sentence, shared by multiple in-house counsel at the Global Legal Forum 2026 in The Hague, captures a shift that is already reshaping how legal departments work with outside counsel. Law firms are adopting AI tools that make their lawyers faster. The question in-house teams are asking: who captures that productivity gain?

If a law firm uses AI to complete a task in 2 hours that used to take 8, and still bills 8 hours, the firm captures all the value. If the in-house team negotiates a project fee instead, both sides share the gain. If the in-house team brings the work in-house entirely using their own AI platform, they capture all of it.

This is not a hypothetical discussion. It is happening now, in every fee negotiation, at every panel review, in every RFP process. The firms that embrace transparency are building stronger relationships. The firms that resist are finding themselves competing on price without goodwill.

This article covers what has changed, what to ask, how to restructure pricing, and when to bring work in-house instead.

What AI Actually Changes About Legal Work Economics

AI tools do not make bad lawyers good. They make good lawyers faster. The economic impact varies by work type:

Work Type	AI Impact	Economic Shift
Document review	80-90% time reduction	Hourly billing model completely broken
Contract drafting	60-80% time reduction for standard contracts	First drafts that took 4 hours now take 30 minutes
Legal research	50-70% time reduction	Research that justified a day of billable time now takes hours
Due diligence	40-60% time reduction	AI-assisted data rooms dramatically reduce review time
Complex negotiation	10-20% time reduction	Still heavily human; AI assists but does not replace
Strategic advice	Minimal time reduction	Value is in judgment, not speed

The pattern: the more routine and document-intensive the work, the greater the AI impact on time. The more judgment-intensive and strategic, the less AI changes the equation.

This means the work that traditionally generated the most billable hours (document review, standard drafting, research) is exactly the work where AI creates the largest gap between effort and billing.

58%

of legal professionals learned about AI through self-exploration, not formal training. Only 19% received any formal AI education.

TED/Global Legal Forum 2026

What to Ask Your Law Firms

Most law firms are using AI in some form. Few are transparent about it. Here are the questions that matter:

1. Which AI tools are you using, and for what?

You have a right to know what tools are being used on your matters. Not because AI is bad, but because it changes what the work is worth and how it should be billed.

Acceptable answers: "We use [tool] for first-pass document review, [tool] for research, and [tool] for draft generation. All output is reviewed by a qualified lawyer before delivery."

Evasive answers: "We have various tools." "That is proprietary." "We are exploring options."

2. How does AI usage affect your billing?

This is the core question. If AI reduces a task from 8 hours to 2, you should not pay for 8 hours. Ask specifically:

Do you adjust hours billed when AI accelerates the work?
Do you have AI-specific billing policies?
Are you open to project-based fees for AI-accelerated work types?

3. What is the human review process?

AI output requires human oversight. Understanding the review process helps you assess quality and fair billing. A firm that uses AI for first drafts and then bills senior associate time for review is delivering a different service than one that uses AI and bills partner review time.

4. What liability do you accept for AI-assisted work?

Most AI providers explicitly disclaim liability. Your law firm should not pass that disclaimer through to you. If the firm delivers AI-assisted work product, the firm's professional liability should cover it the same way it covers any other work product.

Pricing Models That Work

The hourly billing model was designed for a world where time correlated with effort. AI breaks that correlation. Here are the models that replace it:

Project-Based Fees

Fixed fee for a defined scope of work. The firm absorbs the efficiency gain and manages its own costs. You pay for the outcome, not the hours.

Best for: Routine matters with predictable scope (contract reviews, standard transactions, regulatory filings)

How to negotiate: Ask for the fixed fee based on the outcome value, not the estimated hours. If the firm quotes a project fee that equals their estimated hours times their rate, they are just repackaging hourly billing.

Value-Based Pricing

Fee tied to the value delivered, not the time spent. If a firm handles a transaction that saves you $2M in risk exposure, the fee reflects a share of that value rather than the hours invested.

Best for: High-stakes matters where the value is clear and measurable (M&A, complex negotiations, litigation strategy)

How to negotiate: Agree on success metrics upfront. What constitutes value delivered? How is it measured? What is the fee at different outcome levels?

Hybrid Models

Blended approach: fixed fee for routine components, hourly or value-based for complex components. This works when a matter has both predictable and unpredictable elements.

Best for: Multi-phase matters, ongoing advisory relationships, complex transactions with routine sub-tasks

Example structure:

Contract review and due diligence: Fixed project fee (AI-accelerated, predictable scope)
Negotiation support: Hourly with a cap (judgment-intensive, variable scope)
Strategic advice: Retainer with defined access (value is availability and judgment)

Subscription/Retainer Models

Monthly or quarterly fee for defined access and service levels. The firm manages capacity and you get predictable costs.

Best for: Ongoing advisory needs, portfolio management, regulatory monitoring

How to Measure External Counsel Value

Menzies Aviation, as described by their General Counsel Aminata BA at the Global Legal Forum 2026, built a custom AI tool to monitor the value-add of external counsel. This is an emerging practice, but the principles are applicable without custom tooling.

Metrics that matter

Efficiency metrics:

Time from instruction to delivery
Number of revision rounds (fewer is better)
Responsiveness to questions and updates

Quality metrics:

Accuracy of advice (measured by outcomes, not volume)
Relevance of analysis (did they address what you actually needed?)
Proactive risk identification (did they spot issues you did not ask about?)

Value metrics:

Cost per matter type (track over time; should decrease as AI improves)
Outcome relative to cost (risk avoided, value captured, disputes prevented)
Knowledge transfer (did you learn something, or just receive a deliverable?)

Relationship metrics:

Cultural fit with your organization
Understanding of your business context
Willingness to collaborate on pricing and process

The Letterhead Question

Julia Taddei of Petrobras raised an important point at the GLF: "Sometimes you are just paying for the letterhead." External firms commit opinions and put their names behind the work. That has real value, especially for regulatory and board-level matters. Do not discount it. But also do not pay letterhead prices for commodity work.

When to Bring Work In-House

The 10x model for in-house legal teams is not just about handling internal contracts faster. It is also about pulling work back from external counsel that no longer needs to go outside.

Work that should stay external:

Jurisdiction-specific expertise you do not have
Litigation and dispute resolution
Complex regulatory matters requiring specialist knowledge
High-stakes transactions requiring independent opinion

Work that can come in-house with the right platform:

Standard contract drafting (NDAs, service agreements, vendor contracts)
First-pass contract review and redlining
Routine negotiation on standard terms
Contract portfolio management and reporting

The economics: if your law firm charges $400/hour for an associate to draft a standard service agreement (3 hours = $1,200 per contract), and your team can self-serve the same contract in 15 minutes using a playbook-driven platform, the ROI calculation is straightforward.

For more on the self-service operating model, see The 10x Legal Team: Self-Service Contracts at Scale.

20/80

split between legal risk and economic risk in contracts. Legal teams often focus on the 20% while the 80% economic risk goes unmonitored.

Klaas Ehmen, STARCAR / Global Legal Forum 2026

The Conversation to Have Now

Rita Paukste of ILTE Bank summarized what in-house counsel care about from external counsel, in order of priority:

Speed: How fast can you deliver?
Price: What does it cost?
Quality: Is the work good?

AI changes the first two dramatically. Work that was slow is now fast. Work that was expensive should now be cheaper. Quality remains the differentiator.

The conversation to have with your law firms now is not adversarial. It is practical:

"We know you are using AI tools. We support that. We want to understand how it affects the economics of our work together. Let's find a pricing model that reflects the actual effort and value, so we can build a stronger long-term partnership."

The firms that respond well to this conversation are the firms worth keeping. The ones that deflect, resist, or obfuscate are telling you something about how the relationship will evolve.

Frequently Asked Questions

Should we demand discounts because firms use AI?

Not blanket discounts. Instead, restructure pricing to reflect the actual work being done. A 20% discount on hourly rates still leaves the fundamental model broken. Project fees, value-based pricing, or hybrid models are better solutions than discounted hours.

How do we know if our firm is actually using AI on our matters?

Ask directly and include AI transparency requirements in your engagement terms. Some firms are proactively disclosing AI usage. Others will disclose when asked. If a firm refuses to discuss it, that is a red flag about the relationship, not just about AI.

Will pushing on pricing damage our law firm relationships?

No. The firms that take this conversation well are the ones building for the future. Multiple panelists at the Global Legal Forum noted that the direction is clearly toward project-based and value-based pricing. Firms that embrace this proactively build stronger relationships. Firms that resist will find themselves competing on price without goodwill.

How much can we realistically save by bringing routine work in-house?

It depends on your current external spend on routine matters. Organizations that spend $200K-$500K annually on standard contract drafting and review can typically reduce that by 60-80% within the first year of deploying an in-house contract platform. The savings come from eliminating external hourly fees for work that a platform handles in minutes.