Home 01Prompt Builder ✦About Adam 02Books 03Journal 04Contact 05
Home/Journal/Startups
AI for Lawyers Academy · Pick 2026

Paxton AI review: the legal AI startup we recommend to small firms

Most legal AI is built for firms with innovation teams and six-figure budgets. Paxton is one of the few serious platforms designed around the solo lawyer and the small practice. Here is why it earned our recommendation — and the lawyers who should choose something else.

10 min readPricing checked September 2026Editorial recommendation

The verdict

Why it’s our pick

  • One assistant, not five tools: research, drafting and document analysis in a single place — exactly what small firms need.
  • A serious security posture: Paxton states SOC 2, ISO 27001 and HIPAA certification.
  • Built for practice areas small firms serve: personal injury, family, employment, criminal and corporate work, with medical chronologies for PI teams.
  • Honest limits: its research base is US federal and state law, and the monthly price needs real weekly use to justify it.
One-tool coverage
9
Security posture
9
US research depth
8
Ease of starting
8
Non-US usefulness
5

Editorial assessment of fit for US solo and small firms, based on Paxton’s published features, certifications and pricing. Your results should come from your own trial.

What Paxton is

Paxton is an AI legal assistant: a web application that helps lawyers research the law, draft documents and analyse files. Unlike a general chatbot, it is designed specifically around legal work and connected to a legal knowledge base covering US federal regulations, state laws and case law across all 50 states.

Paxton says thousands of lawyers use it, and that its clients include enterprise legal teams such as Bank of America. But its most distinctive quality is where it chose to focus: independent lawyers and smaller firms, who have historically been an afterthought for legal technology vendors.

The company behind it

For a tool that will hold client documents, who built it matters. Paxton was founded by Tanguy Chau (CEO) and Michael Ulin (CTO). Chau holds an engineering Ph.D. and an MBA from MIT, spent seven years as a Silicon Valley venture capitalist, worked as an engagement manager at McKinsey, and served as chief compliance officer at a FINRA- and SEC-regulated investment adviser — a background that shows in the company’s emphasis on security and verification.

The company raised a $6 million seed round led by WVV Capital, with participation from Kyber Knight and 25Madison, followed by a $22 million Series A. That funding is not a guarantee of anything, but it does suggest the product is likely to be maintained and improved — an important consideration when choosing a vendor for the long term.

What it does

FeatureWhat it helps with
Legal researchContextual answers drawing on US federal and state law and case law, for issue-spotting and early research.
DraftingFirst drafts of legal documents, clauses, emails and contracts.
Document analysisUploading files for summaries, insights and recommended next steps.
Medical chronologiesOrganising medical records into timelines — valuable for personal injury practices.
Billing summariesSummarising medical billing for damages work.

The practical value is consolidation. A small firm might otherwise pay separately for a general AI assistant, a research add-on and a document-review tool — and still have to move material between them.

Security and accuracy

Paxton states that it is SOC 2, ISO 27001 and HIPAA certified and operates as a secure, closed model. For lawyers handling medical records and sensitive client files, those independent attestations are exactly what you should look for. As with any vendor, read the data-use and retention terms in the agreement itself before uploading client material.

On accuracy, Paxton reports a 94% non-hallucination rate on the Stanford Legal Hallucination Benchmark. Treat vendor-reported benchmarks as a useful signal rather than a promise. A 94% rate still means errors can occur — and courts have sanctioned lawyers who relied on legal-specific tools without checking. Verification remains your job; our 7-step verification checklist works for every tool, Paxton included.

Our standard for every recommendation: no AI tool is a substitute for reading the authority. We recommend tools that make a careful lawyer faster — never ones that promise you can stop being careful.

Pricing

At the time of writing, Paxton’s published pricing is:

PlanPriceNotes
Individual — monthly$499 per user / month7-day free trial
Individual — annual$2,999 per user / yearPaxton describes this as a 50% saving versus monthly
EnterpriseCustomFirm-wide access, onboarding, collaboration, priority support

The annual plan works out at roughly $250 a month. If it saves a lawyer even two or three billable hours a month, the arithmetic is straightforward. The monthly plan makes more sense as an extended trial than as a long-term arrangement. Prices change — confirm on Paxton’s pricing page.

Who should — and shouldn’t — use it

A strong fitConsider something else
US solo lawyers and small firms wanting one AI assistantLawyers whose work is mainly outside the US
Personal injury teams handling medical recordsTransactional lawyers who work almost entirely in Word (Spellbook may suit better)
Family, employment, criminal and corporate practitionersFirms already paying for Lexis+ AI or CoCounsel research
Firms that need documented security certificationsLawyers who only need occasional drafting help (a business-plan general assistant is cheaper)

Getting real value from a 7-day trial

Seven days is short. Plan it before you sign up:

  1. Day 1: choose three real, recently completed matters — so you already know the right answers.
  2. Days 2–3: run the research questions from those matters and compare Paxton’s results with what you found.
  3. Day 4: ask it to draft a document you have already drafted, then compare quality and editing time.
  4. Day 5: upload an anonymised or non-confidential file for analysis and check every point against the source.
  5. Days 6–7: estimate hours saved per week, review the contract terms, and decide.
AI for Lawyers 2026 book cover
Whichever tool you choose

AI for Lawyers 2026

Prompting methods and 100 practice-area prompts that make any legal AI assistant more useful.

See it on Amazon

Alternatives worth comparing

  • Clio with Vincent AI — best if your firm already runs on Clio.
  • Lexis+ AI or CoCounsel — for research-heavy practices already invested in LexisNexis or Westlaw.
  • Spellbook — for contract drafting inside Word.
  • Claude or ChatGPT on a business plan — for general drafting at lower cost, paired with a separate research tool. See Claude vs ChatGPT for lawyers.

About this recommendation: “Academy Pick” is an editorial designation. It reflects our view of how well a product fits the lawyers we teach, based on published information. We update picks when products, pricing or our view change.

Frequently asked questions

How much does Paxton AI cost?

Paxton lists $499 per user per month, or $2,999 per user per year on the annual plan, with a 7-day free trial. Enterprise pricing is custom.

Is Paxton secure enough for client data?

Paxton states SOC 2, ISO 27001 and HIPAA certification. Review data-use and retention terms in the agreement before uploading client material.

Does Paxton hallucinate?

Paxton reports a 94% non-hallucination rate on the Stanford Legal Hallucination Benchmark. That is a strong figure, but not zero — verify every authority before relying on it.

Can lawyers outside the US use Paxton?

Drafting and document analysis can help anywhere, but its research knowledge base is US law. Non-US lawyers should test it against their own jurisdiction first.

Choosing tools is easier with training

We help lawyers and firms pick, test and adopt AI safely.