Back to Blog

A Series B startup uploads a vendor agreement to Claude. Four minutes later, it has a fully redlined contract with 19 strategic edits, margin comments explaining each change, suggested alternative language, and a negotiation strategy prioritised by business impact.

That's not a demo. That's an actual use case showcased on Anthropic's own resources page. And it should terrify every mid-tier law firm in the country.

What Claude Actually Does With Contracts

Let's be precise about what's happening here, because the legal industry has been remarkably successful at hand-waving AI capability as "just a tool."

Claude reads the entire contract. It understands the business context — that this is a startup with limited budget flexibility negotiating with a larger vendor. It then reviews every clause through a commercial lens, identifies risks, and produces tracked-change redlines with explanatory comments. Specifically, it flags and edits auto-renewal traps, liability exposure, data ownership gaps, and termination inflexibility.

It then generates a comparison table of original versus suggested terms, a prioritised negotiation strategy, and alternative clause language calibrated to the power dynamic between the parties.

This isn't pattern matching. This is substantive contract analysis. And Claude delivers it for the cost of a monthly subscription that wouldn't cover 12 minutes of a commercial solicitor's time.

The Billable Hour Is a Walking Corpse

The legal profession's business model — particularly in commercial, corporate, and transactional work — is built on time. Specifically, on the time it takes smart humans to read documents carefully and apply judgement. Claude's 12x speedup (14.8 minutes versus 3.8 hours) doesn't just dent that model. It collapses the economic foundation.

Consider a typical contract review engagement. A mid-level associate at a City firm bills at £350–500 per hour. A standard vendor agreement review takes 3–5 hours. That's £1,050–2,500 per contract, before partner oversight and administrative costs.

Claude does substantively the same first-pass work in minutes. Not cheaper. Not "more efficient." Categorically different in cost structure.

The firms that survive this won't be the ones that ignore it. They'll be the ones that rebuild their value proposition around the 20% of legal work that requires genuine human judgement — the courtroom advocacy, the complex multi-party negotiations, the regulatory strategy that requires reading political rooms — and use Claude for everything else.

"But AI Can't Replace Legal Judgement"

This is the defence the profession is clinging to, and it's both correct and irrelevant.

It's correct because Claude doesn't replace the senior partner's instinct about which battles to fight in a negotiation, or the litigator's read on a judge's temperament, or the regulatory lawyer's understanding of where enforcement priorities are shifting.

It's irrelevant because that's maybe 15–20% of what law firms actually bill for. The other 80% — document review, due diligence, contract drafting, compliance checking, research memoranda — is exactly the kind of structured, text-heavy, pattern-recognisable work that Claude excels at.

The legal industry isn't facing an existential threat to judgement. It's facing an existential threat to the business model that subsidises judgement with billable grunt work.

The Access to Justice Angle Nobody Wants to Discuss

Here's where it gets properly controversial: Claude's contract review capability is arguably better for justice than the current system.

Right now, a small business signing a vendor agreement has two options: pay £2,000+ for a proper legal review, or sign it unreviewed and hope for the best. Most choose option two. They get locked into punishing terms, unknowingly waive important rights, and only discover the problems when it's expensive to fix them.

Claude gives that small business institutional-quality contract analysis for £15 a month. It flags the auto-renewal traps, the unlimited liability clauses, the intellectual property assignments buried in Schedule 3. It doesn't just level the playing field — it democratises a capability that was previously gated behind a price point most businesses couldn't justify.

If we're being honest, the legal profession's discomfort with AI isn't primarily about quality concerns. It's about the uncomfortable reality that high-quality legal analysis shouldn't cost what it currently costs, and the technology now exists to prove it.

What the Smart Firms Are Doing

The forward-thinking firms — and they exist, though they're quieter about it than you'd expect — are doing three things.

First, they're using Claude internally and not telling clients. They're maintaining billing rates while dramatically reducing the actual time spent on document-heavy work. This is ethically questionable and economically rational, which means it's widespread.

Second, they're repositioning as "AI-augmented" practices, using Claude for first-pass work and billing for the human review and strategic overlay. This is more honest and more sustainable, but requires accepting significantly lower revenue per engagement.

Third, they're building productised legal services — fixed-fee contract review packages, subscription compliance monitoring, automated due diligence reports — that use Claude as the engine and human lawyers as the quality control layer. This is the model that will win.

The Timeline

Claude's autonomous capability — the number of consecutive actions it can take without human intervention — has doubled in six months. Contract review is among the most structured, well-defined tasks in professional services. It will be among the first to reach near-full automation.

Within 18 months, AI-assisted contract review will be the default for SMEs. Within three years, it will be the default for mid-market transactions. The firms still billing 5 hours for a first-pass contract review in 2028 won't be premium. They'll be extinct.

Professional services firm navigating AI disruption?

Digital by Default helps professional services firms navigate AI adoption without destroying their value proposition. It's a narrow path, but it exists.

Get in Touch →