Will AI Replace Lawyers? 85% Already Use It, and Most of the Work Is Still Manual

Via GlobeNewswire: 85% of Lawyers Use AI, Yet Manual Work Still Dominates Legal Workflows, New Progress Software Report Reveals Updated October 23, 2026 with what the data says about AI replacing lawyers, real examples, what a small firm should automate first and an FAQ.

Your associates have ChatGPT. Your intake queue still has a fax machine energy.

Progress Software's State of Legal 2026 benchmarking report, based on a nationwide survey of 304 U.S. lawyers conducted by Regina Corso Consulting, finds that 85% already use AI for legal research, document summarization, and intake support. That is not the headline. The headline is what happens next: 77% say much of their work remains manual, and 73% report workflows with too many steps.

Loren Jarrett, EVP and GM of Digital Experience at Progress, put it bluntly: firms are layering AI on top of broken processes instead of rethinking how work gets done. For solo and small law firm owners, that is the difference between buying another subscription and reclaiming billable hours that never make it to the invoice.

Note: This report was produced by Progress Software, which sells legal workflow products. The adoption-without-redesign gap aligns with independent trade coverage elsewhere in professional services — but treat vendor benchmarks as directional, not gospel.

So, will AI replace lawyers?

Not the lawyer. Parts of the job, yes. The evidence points to AI taking over research, summaries and intake, and putting pressure on the billable hour, while judgment, advice and accountability stay with people.

• Use is high, automation is low. In the Progress survey, 85% use AI, but only 24% say a significant portion of their daily work is automated. (Progress Software release)
• The biggest forecast is about tasks, not jobs. Deloitte Legal projects that within three to five years AI agents could handle about 30% of the work inside corporate legal teams, and about 20% of in-house lawyers could work as hybrid engineer-lawyers. That is a projection for in-house teams, not law firms. (Artificial Lawyer on Deloitte; our take: what Deloitte's 30% means.)
• Billing is where the change shows first. In Deloitte's survey, 85% of respondents expect AI to change how law firms price work, with hourly-rate work expected to fall from 72% to 44% within two to three years.

One lawyer quoted in the Progress release put it this way: “AI is not here to take our jobs as lawyers. If we can figure out how to harness it, we can free ourselves up to deliver incredible value.” — Whitney Harper, co-founder, ADVOS legal.

The risk for a firm is not being replaced by software. It is a competitor doing the same work in fewer hours and pricing it as a fixed fee.

What AI already does in legal work

Contract review and intake at Cebu Pacific. The airline's legal team cut contract review from an average of 2.5 hours to around 1.2 hours per pass using ChatGPT Enterprise. Project intake and triage fell from five days to around one day. People still make the calls. The AI does the first pass. (TechNode Global; our write-up: Cebu Pacific's legal review.)

Big firms running AI on their own servers. Latham & Watkins bought Nvidia GPU servers to run AI in-house so confidential client work stays inside the firm. It shows a large firm treating AI as core infrastructure for legal work, not a replacement for its lawyers. See why one of the world's biggest law firms runs AI on its own servers.

Small firms using AI without the payoff. Clio's 2026 data, summarised by Above the Law, shows about 71% of solos and 75% of small firms use AI, but only about a third have increased revenue with it. 86% of solo firms and 78% of small firms have not changed their pricing for AI. More in why solo firms adopted AI but can't turn it into revenue.

What a small law firm should automate first

Start where the delay is measurable and the work is repeatable. Keep judgment calls with lawyers.

1. Client intake. 47% of lawyers in the Progress survey say intake takes four days or longer, and 92% say automation would let them take on more business. Map every step from first contact to opened matter, then automate the form, the data entry and the reminders.
2. Document summaries and first-pass review. Summaries of long documents and first-pass contract review, with a lawyer checking every output. Cebu Pacific's numbers show where the time goes.
3. Status updates and routine client emails. Drafted by AI from the matter file, sent by a person.
4. Research memos. A first draft to check, never a final answer. Every citation gets verified.
5. Then fix the pricing. If a task drops from five hours to one, move it to a fixed fee before the saved hours turn into a discount.

Before any of it, write a one-page AI policy: which tools are approved, what client data may go into them and who reviews the output. For how we map legal workflows, see our law firm page.

Automation enthusiasm, automation reality

The survey reports a familiar split. 82% of lawyers say automation has improved their work lives — yet only 24% say a significant portion of daily work is actually automated. Integration challenges, budget constraints, and security concerns remain the top barriers.

Client intake is where the friction shows up in dollars. Nearly half (47%) say intake takes four days or longer, even though most believe it should take two to three. 94% say faster intake would improve outcomes; 92% say automation would let them take on more business. The demand is there. The plumbing is not.

Meanwhile, 52% describe themselves as effective but not efficient — manual tasks, unnecessary handoffs, fragmented systems. 84% say inconsistent processes across teams reduce efficiency. 36% cite lack of governance or training as adoption scales. AI without standards is just faster inconsistency.

What smart firms do before the next tool demo

Owner-led law firms do not need a $2.5 billion embedded-engineering program. They need fewer steps between "client called" and "matter opened." Practical moves:

• Map intake end-to-end — count handoffs from web form to conflict check to engagement letter; that is where four-day delays hide.
• Standardize before you automate — if three partners run three versions of the same checklist, AI will summarize three flavors of chaos.
• Pair AI with governance — acceptable-use rules, training, and a named owner so associates know which tools may touch client material.
• Measure cycle time, not licenses — days to open a matter beats "we have Copilot seats."
• Pilot one workflow — research memos or intake triage first; prove hours returned before firm-wide rollout.

Progress's survey found 95% want legal technology that is simple and intuitive. Complexity is not a strategy. Neither is another dashboard nobody opens after month two.

How BuildBrain finds the hours AI was supposed to free

BuildBrain is a vendor-neutral managed AI services provider for owner-led SMEs — including law firms too small for enterprise transformation theater.

Start with a Workflow ROI Audit. We trace where expensive staff time actually goes: intake, research, drafting, status updates, client follow-up. We separate tasks AI should own from tasks that need partner judgment — and cost the roadmap so you can kill tools that speed up the wrong work. Fixed-fee audits typically run $2,500–$7,500.

If associates may already be pasting matter details into consumer AI accounts, pair the audit with a Shadow-AI Risk Assessment so process redesign and confidentiality controls move together.

Managed AI Operations is the destination: ongoing monitoring, spend controls, and reporting that ties AI to billable outcomes — not shelfware.

See audit options or request an assessment before the next renewal buys speed without intake fixes.

Adoption is not redesign

Eighty-five percent AI usage is a vanity metric if seventy-seven percent of the work still walks through manual doors. Owner-led firms that want AI to pay for itself should eliminate friction first — intake, handoffs, single source of truth — then let the models do what they are good at.

That is less exciting than announcing "we are an AI-forward firm." It is also how you stop paying twice: once for the tool, once for the rework.

Start with a Workflow ROI Audit. Find out where AI saves real time — and where it is just decorating a 1997 workflow.

Frequently asked questions

Will AI replace lawyers?

Not on current evidence. AI is taking over tasks such as research, document summaries and intake. Lawyers still give the advice, make the judgment calls and carry the responsibility. The bigger change is to billing: Deloitte's survey expects hourly-rate work to fall from 72% to 44% within two to three years.

Will AI replace paralegals?

Paralegal work includes many tasks AI already does well, such as summarising documents, organising files and drafting routine letters, so more of the role is exposed than a lawyer's. Paralegals who check AI output, manage matters and handle clients remain needed.

Will AI take over law jobs?

It is taking over parts of them. Deloitte Legal projects AI agents could handle about 30% of the work inside corporate legal teams within three to five years. That is a share of tasks, not of jobs, and it is a projection.

Will AI replace attorneys in the next 10 years?

No reliable source says so. The forecasts that exist look three to five years ahead and describe AI handling a share of the work. The more likely change is that attorneys who use AI well win work from those who don't.

What should a small law firm automate first?

Client intake first, because it is slow and measurable: 47% of lawyers say it takes four days or longer. Then document summaries, first-pass review, status updates and research drafts, each checked by a lawyer. Then adjust pricing so saved hours don't become a discount.

This article summarizes publicly reported information and is for general informational purposes only. It does not constitute legal, tax, financial, investment, security, or compliance advice. BuildBrain is not a law firm, accounting firm, or registered investment adviser. Facts, pricing, statistics, and product capabilities cited here reflect the sources listed at the time of writing and may change. Readers should verify current information independently and consult qualified professionals regarding obligations specific to their industry, jurisdiction, and circumstances—including applicable federal, state and local requirements. BuildBrain may have commercial relationships with vendors mentioned; where material, such relationships are disclosed. Nothing in this article is an endorsement of any specific AI product, model, or provider.