Harvey raises $160M to push AI deeper into legal workflows
Harvey AI Corp. closed a $160 million Series F led by Andreessen Horowitz, with new participation from WndrCo and accounts advised by T. Rowe Price Associates Inc., and returning backers Sequoia, Kleiner Perkins, Conviction, and Elad Gil. The round values the company at $8 billion, up from $5 billion after a $300 million raise in June, and follows a recent €50 million (about $58 million) investment from EQT Growth.
The pitch is simple: give attorneys AI that actually helps with complex legal work. Harvey says more than 700 firms and enterprises use its platform, with over 74,000 attorneys on board. Notable customers include Allen & Overy and PricewaterhouseCoopers LLP.
Why this matters for law firms and in-house teams
Harvey's tools support contract analysis, due diligence, compliance, and litigation work. Think faster issue spotting, consistent playbook application, and cleaner audit trails - without forcing teams to adopt a new way of working.
For in-house counsel, the upside is shorter deal cycles and clearer collaboration with outside firms. For law firms, it's a way to protect margin on fixed-fee matters and standardize quality across teams.
Under the hood: domain-trained models, not just search
Harvey uses multiple AI models, including a custom-trained model built with OpenAI for case law. That focus on legal domain knowledge supports multi-step workflows instead of single-answer chat.
"If you just do retrieval, you can answer very simple questions about areas of law that you aren't really an expert in, but that's actually not that useful for most attorneys," said Winston Weinberg, Harvey's co-founder and CEO. "With case law research, you're finding ammo for your argument, and that's much more difficult to do."
Harvey also partnered with PwC to co-build a tax AI assistant that combines curated tax datasets with Harvey's LLM knowledge. That's a blueprint for more verticalized assistants in specialized practice areas.
Shared Spaces: secure collaboration without giving up your playbook
Alongside the funding, Harvey introduced Shared Spaces, a secure, AI-driven workspace built for attorneys. Firms can share customized tools - workflows, playbooks, and checklists - without exposing proprietary prompts or the logic behind them.
Shared Spaces brings documents, analysis, and workflows together so teams work side-by-side instead of losing context in email threads and version control. "We all know that legal work is deeply collaborative - it's about partnership," said Rick Liu, strategic business development lead at Harvey. "And that partnership deserves a way to work that builds on that collaboration, instead of getting lost in email chains and version control."
Spaces provides access to Harvey's models, workflow builders, and AI agents. You can invite non-Harvey users, assign roles and permissions, and audit who accessed, edited, or generated each artifact.
Where Harvey fits today: practical legal use cases
- Contract review: clause extraction, playbook alignment, fallback suggestions, and exceptions logs that tie back to negotiation history.
- Due diligence: automated summaries across deal rooms with cross-references, risk flags, and a crisp binder for counsel and the business.
- Litigation support: case law research packs with quotes, citations, and draft argument outlines that can be verified and expanded by attorneys.
- Compliance operations: policy comparisons, control mapping, and audit-ready trails for updates and approvals.
- Tax: the PwC co-built assistant for faster issue identification and documentation across complex tax questions.
Governance and risk questions to pin down
- Data boundaries: where data is stored, who can access it, and how client matter segregation is enforced.
- Prompt and artifact security: how Spaces permissions protect playbooks, prompts, and outputs across matters and clients.
- Auditability: exportable logs for internal audit, QA, and outside review; alignment with your SOC 2/ISO workflows.
- Model behavior: source transparency, citation reliability, testing methods, and procedures for error handling and conflicts.
- Cost and control: usage caps, cost centers by matter/client, and safeguards for sensitive or privileged content.
What the funding signals
Series F at an $8 billion valuation points to scale: more model depth, more practice-specific workflows, and bigger enterprise rollouts. Expect tighter integrations with document systems, DMS, email, and e-billing to reduce swivel-chair friction.
The investor mix - a16z, Sequoia, Kleiner Perkins, and T. Rowe Price - suggests pressure to prove measurable outcomes: cycle-time reduction, lower review costs, and fewer errors under scrutiny. That's good for buyers who need ROI clarity before standardizing.
How to pilot Shared Spaces and show value quickly
- Pick two repeatable workflows (e.g., vendor NDAs and DPAs) and define success metrics: turnaround time, redline volume, outside counsel spend, and error rates.
- Lock down governance early: roles, permissions, data retention, and audit exports mapped to your policies.
- Create a verification loop: attorney review, citation checks, and sampling standards so outputs remain reliable.
- Publish a simple playbook: when to use AI, what to verify, and how to escalate issues - then iterate every two weeks.
- Report outcomes to leadership monthly: time saved, cost impact, and quality signals that matter to the business.
Bottom line
Harvey is pushing deeper into legal work that demands reasoning, citation quality, and collaboration across firms and in-house teams. Shared Spaces is a practical step for security-conscious teams that want speed without giving up control.
If your legal org is planning structured upskilling around AI, see AI courses by job roles here: Complete AI Training - Courses by Job.
Your membership also unlocks:
