Legal Tech / Drafting & general assistants
Harvey
Harvey AI (Counsel AI Corporation) · HQ US · founded 2022 · Last checked 2026-09-04
Last checked = the date we last re-opened the vendor's published pages. Facts are what the vendor documents, not what it claims in marketing.
Firm-wide legal AI platform for research, drafting, review and agentic workflows, deployed by 40+ Am Law 100 firms and in-house teams; first legal AI agents certified under AIUC-1 (July 2026).
- Category
- Drafting & general assistants · research, drafting and review, agentic workflows, Word add-in
- Built for
- Large firms, Mid-sized firms, In-house legal
- Model providers (disclosed)
- OpenAI, Anthropic, Google, Microsoft Azure, Mistral, Baseten, Fireworks, AWS
- Data residency options
- US, EU, CH, AU
- Pricing
- enterprise
- Markets
- US, UK, EU, AUSTRALIA
- Last checked
- 04 Sept 2026
What the vendor documents
Recorded from the vendor's own trust, privacy and terms pages. "Not disclosed" means we found no public statement, not that the control is absent.
What the public documentation states
Harvey’s public pages describe a legal AI platform for research, drafting, review and agentic workflows used by law firms and in-house teams. The sub-processor list dated 2 August 2026 names Microsoft, OpenAI, Google, AWS, Anthropic, ElevenLabs, DeepL, Mistral, Baseten and Fireworks, and the security page lists EU or Switzerland, US, and Australia residency options. The security page states that Harvey does not use inputs, outputs or uploaded documents to train underlying models and requires Zero Data Retention from model providers. The Security Addendum dated 26 December 2025 states a 48-hour notice after becoming aware of a security incident and one-year audit-log retention. The security page lists SOC 2 Type II, ISO 27001, ISO 27701 and ISO 42001. The DPA dated 2 August 2026 says customer content is deleted within 30 days after termination. Public pages read for the audit did not publish pricing or a standing accuracy figure; separate record sources list VLAIR and AIUC-1 evidence.
In the vendor's words
“We don’t use inputs, outputs, or uploaded documents to train underlying models.”
“Harvey requires Zero Data Retention (ZDR) by model providers.”
“within 48 hours after becoming aware”
“SOC 2 Type II”
Not disclosed on public pages
"Not disclosed" records that no public statement was found where we looked, on the date shown. It does not assert that the control, term or figure is absent.
On pages read on 2026-09-06, Harvey did not publish a list price, trial, minimum term, seat minimum, private-deployment option, standing accuracy claim or accuracy method on the checked vendor pages. A procurement or DPIA file normally needs pricing and seat scope for spend approval, deployment options and residency for client confidentiality, and accuracy method details for SRA or ABA competence and supervision controls.
Questions to put to the vendor
- Which of Microsoft, OpenAI, Google, AWS, Anthropic, ElevenLabs, DeepL, Mistral, Baseten and Fireworks process matter data for the selected workflow?
- For an EU or Switzerland residency selection, which inference providers and regions are used?
- Does any private-deployment option exist for Harvey agents, or only tenant-level configuration in shared cloud services?
- What current accuracy evaluation method does Harvey use outside VLAIR and BigLaw Bench?
- Does the 48-hour notice run from first awareness by Harvey or from confirmed incident classification?
Accuracy evidence
- third-partyVals AI Legal AI Report (VLAIR), 27 Feb 2025 Opted into 6 of 7 tasks; matched or outperformed the lawyer baseline in five and outperformed the other AI tools in four (e.g. Document Q&A 94.8% vs 70.1% baseline); below the lawyer baseline on redlining (65.0% vs 79.7%).
- vendorHarvey BigLaw Bench (vendor-run), Aug 2024 74% answer score on vendor-designed tasks.
- vendorAIUC-1 AI agent certification, July 2026 Third-party certification scheme announced by the vendor; scope covers Harvey agents.
Related incidents
No incidents are recorded for this tool as of 2026-09-04.
Ownership · Venture-backed; $200M Series G at an $11B valuation (Mar 2026, GIC and Sequoia); ~$1.22B raised.
News and analysis about Harvey →
Sources
- Harvey — Security
- Harvey — Security addendum
- Harvey Data Processing Addendum (last updated 2 August 2026)
- Harvey sub-processor list (last updated 2 August 2026): OpenAI, Anthropic, Google, Microsoft, Mistral, Baseten, Fireworks; regions US, EU, Switzerland, Australia
- Harvey — Security by design
- Harvey — AIUC-1 certification
- Vals AI — VLAIR report, Feb 2025
- CNBC — Harvey $200M at $11B (Mar 2026)
- SafeLegalAI documentation audit, 2026-09-06: every fact with its quote, source page and archived copy
Report an issue with this record →
Cite this record
SafeLegalAI Legal Tech Tools, "Harvey", safelegalai.com/tools/harvey (last checked 2026-09-04, accessed 2026-10-07). Data: CC BY 4.0.
What the vendor has announced, dated
- Harvey partners with Everlaw · Partnership
- Harvey rebuilds Playbook Review as multi-agent system · Feature release
- Harvey introduces Horizon Scanning · Feature release
- Harvey makes Fable 5.1 available · Model or provider disclosed
- Harvey adds Memory · Feature release
- Harvey shares Tenet research preview · Model or provider disclosed
- Harvey introduces Harvey II · Product launch
- Harvey reports frontier review-table model training · Research or benchmark
- Harvey adds audio, video and image understanding · Feature release
- Harvey publishes August product updates · Feature release
- Harvey updates Outlook experience · Feature release
- Harvey adds Playbook Builder · Feature release
First-party statements from the vendor's own pages, as dated; "announced" is not "in use today". This month across all vendors · release timeline.
Documented facts, side by side
- Harvey and Lexis+ AI (with Protégé)
- Harvey and Westlaw Precision / CoCounsel Legal
- Harvey and Microsoft 365 Copilot
- Harvey and ChatGPT Enterprise / Team
- Harvey and Claude for Enterprise / Team
- Harvey and DeepIP
- Harvey and Definely
- Harvey and DraftWise
- Harvey and Legora
- Harvey and Wordsmith
Facts from both vendors' pages in one table; never a ranking. All comparisons.
Also in Drafting & general assistants
SafeLegalAI is a research publication, not a law firm. Nothing here is legal advice, and no lawyer–client relationship arises from reading it. Rules change; always check the official document linked on each record and take advice on your own situation. Researched and drafted with AI assistance; verified against primary sources and edited by SafeLegalAI. The linked official documents are the record — our summaries are not the law and are not legal advice.
Vendor, product, court and regulator names and logos are the property of their respective owners and are used solely to identify the product, body or document referred to; SafeLegalAI claims no rights in them. SafeLegalAI is not affiliated with, endorsed by, sponsored by or approved by any listed vendor, except where a record is labelled an affiliated product. Facts are recorded from vendors' published pages on the date shown; "not disclosed" means not found on the pages read, not that the vendor lacks it. No ranking, rating, review, recommendation or warranty is implied. Mark owners who prefer a text-only reference, and vendors seeking a correction, may write to the report page. Trade mark and quotation policy.
