What it is
GenieAI is a London contract AI built for companies that draft and review their own contracts without a big legal department. It started in 2017 as a legal-drafting research project (founders Rafie Faruq and Nitish Mutha, out of Entrepreneur First), built a free open-source template library, and today sells “Genie”: a web app that drafts, reviews, redlines, researches and negotiates contracts inside your own Word files, with an add-in that brings the assistant into Microsoft Word. It raised a $17.8M Series A led by GV with Khosla Ventures in October 2024 and reports more than 280,000 users. That figure is self-reported and counts users, not paying customers.
Two launches in 2026 changed what you are buying. Genie 3.0 (public trial from May 2026) added “Company Knowledge”: you upload past contracts, policies and negotiation history by department, and review runs against your own positions rather than a generic market view. On 5 November 2026 GenieAI shipped Agentic Legal Specialists, where a lead agent plans each task and assigns drafting, negotiation, risk, research and proofreading agents with quality gates between them. The vendor’s own example is 2 agents and 1 gate for a two-page NDA, 12 agents and 3 gates for a 40-document due diligence. The benchmarks GenieAI publishes against ChatGPT and Claude Cowork are vendor-run; treat them as a reason to pilot, not as evidence.
What it costs in practice
GenieAI is one of the few contract AI tools in this segment that still publishes prices (USD, checked 20 November 2026):
- Free: $0, one user, enough AI tokens for about one task, no document exports and no PDF upload. It is a demo, not a working tier.
- Pro: $75/month, or $59/month billed annually ($708/year). One user, about 5 documents a month on the included tokens, top-ups available.
- Business: $320/month, or $259/month billed annually ($3,108/year) for up to 5 users, about $52 per user. About 50 documents a month shared across the team, plus the Word add-in, integrations, team playbooks and the contract repository.
- Enterprise: quoted, billed annually, with unlimited tokens, role-based department spaces, SLA and a pilot on your own contracts.
The real-world band for a small in-house team is therefore $3,100-3,900 a year on Business, against quote-only pricing from Spellbook, where third-party estimates for the entry seat range from roughly $20 to $99 per user per month and nobody outside its sales team knows the real number. The catch is that the plans meter AI tokens, not seats. “About 5 documents a month” on Pro means a founder reviewing one or two contracts a week will hit the cap.
Best for
- Founders, finance leads and first in-house lawyers at UK and EU SMEs who draft NDAs, MSAs, employment contracts and supplier terms themselves and want what one G2 reviewer called a draft about 90% ready for a lawyer to check
- Two-to-five-person legal or commercial teams that want playbook-based review of incoming paper at a published price, with a Word add-in and DocuSign or Adobe Acrobat Sign in the same tool
Watch-outs
- The Word add-in and integrations sit on Business, not Pro. On Pro you work in GenieAI’s web app and export to Word. Guard: if your lawyers live in Word, budget Business from day one and pilot the add-in on a real redline before committing.
- Your data’s treatment depends on the plan. Zero data retention with OpenAI and Anthropic is listed for Business and above only, and the pricing FAQ says you opt out of AI training by emailing help@genieai.co, which implies it is on until you do. Guard: send the opt-out request and get the DPA signed before uploading any client or counterparty contract.
- ISO 27001, not SOC 2. GenieAI lists ISO 27001 and UK/EU hosting on AWS; SOC 2 does not appear anywhere on its site. Guard: if a US customer or your own security team requires a SOC 2 report, ask for it before the trial, not after.
- The agentic layer is weeks old. Legal Specialist Agents launched on 5 November 2026 and the Genie MCP server for Claude and ChatGPT is marked “coming soon”. Guard: buy on the drafting and review that work today and pilot multi-document work on a closed matter before relying on it.
Alternatives and when to pick them instead
Spellbook is the most widely adopted Word-native contract AI for lawyers (4,500+ legal teams by its own count) and wins on Word-native depth and precedent search across iManage; pick it if your users are practising lawyers redlining all day and you can accept a sales-led quote. Juro is the other default for SMEs and scale-ups and wins when the problem is the contract lifecycle (approvals, signature, repository, renewals) rather than drafting quality. Legora is the fastest-growing entrant in legal AI; test it if you are a firm or a large department that wants a collaborative agent workspace and a research layer, not a published SME price. For a comparison of the drafting-first options, see alternatives to Spellbook.
If you only need clean first drafts of standard documents and never review counterparty paper, GenieAI’s free open-source template library may be all you need; the Free plan, with no exports, will not be. Pay for Pro only once you review more than a contract or two a month.