SAN FRANCISCO, February 12, 2025 — Harvey today announced a $300 million Series D led by Sequoia, valuing the legal and professional-services AI company at $3 billion. The financing will support product development, international expansion and teams that help organisations introduce AI into governed professional work.

The company announcement describes a platform intended for research, drafting, analysis and other knowledge-intensive workflows. Harvey is designed to assist qualified professionals; it does not assume their legal duties or provide an independent substitute for professional judgment.

New capital follows adoption in professional services

Harvey develops generative-AI tools for law firms, corporate legal departments and other professional-services organisations. The platform combines general model capability with customer knowledge, matter context and workflow controls. Users can investigate a question, analyse document sets and prepare work products inside an environment configured for professional use.

The $300 million Series D is private growth financing rather than customer revenue. Bloomberg reported a $3 billion valuation for the February round. That figure reflects negotiated investment terms and should remain separate from operating measures such as recurring revenue, active users, retention and the quality of completed work.

Blank case papers and folders pass from a transparent secure archive through an unmarked computation block to three clean source cards and a brass human-verification gate
The visual separates protected source material, machine processing, retrieved evidence and professional review instead of treating one generated answer as the whole service.

Areas supported by the financing

  • research, drafting and document-analysis products;
  • secure connections to approved organisational knowledge;
  • implementation and customer-success teams;
  • international operations and locally relevant workflows;
  • evaluation, governance and administrative controls.

Sequoia leads a continuing investor group

Sequoia led the round. Existing investors including the OpenAI Startup Fund, Kleiner Perkins, Elad Gil and Coatue also participated, according to Harvey. Their continued backing gives the company resources to build across product, infrastructure and customer deployment.

Funding does not remove the need for capital discipline. Serving professional organisations involves model and cloud costs, security work, integrations, specialist support and sales cycles. Sustainable growth depends on whether recurring customer value develops faster than the full cost of delivering and governing the service.

Legal work requires evidence, not merely fluent language

A legal answer can affect rights, transactions and disputes. Useful assistance therefore requires identifiable sources, correct jurisdiction, current law and a record of how evidence was selected. Fluent output without those controls can conceal a missing authority or an assumption that does not fit the matter.

Harvey's product work will continue to emphasise retrieval, citations and document context. Customers remain responsible for deciding which sources are authoritative, checking the output and determining whether it can be used in advice, a filing, a contract or an internal decision.

Confidentiality begins with matter boundaries

Professional-service organisations hold privileged and commercially sensitive information. Access must follow the client, matter, team and role rather than a general organisational permission. Logs, retention rules, encryption and model-data controls must be understandable to security teams and practising professionals.

Expansion into more organisations and countries increases this responsibility. Data-location requirements, professional rules and expectations about secrecy differ across jurisdictions. Product configuration and customer agreements must reflect those differences rather than treating global availability as one uniform deployment.

Implementation is an operating change

Introducing an AI workspace is not completed when user accounts are created. Firms need approved use cases, training, source libraries, escalation rules and measurements that distinguish experimentation from dependable production work. Leaders must decide which tasks can be accelerated and which require tighter review.

Useful measures include time saved on a defined workflow, citation accuracy, correction rates, adoption by practice group and the effect on client service. Raw prompt counts do not reveal whether the organisation improved quality or economics.

Professional responsibility remains with people

Lawyers and other regulated professionals continue to owe duties of competence, confidentiality and supervision. They must understand the limitations of the tools they use and review material in proportion to its risk. A system can organise evidence and prepare a draft; it cannot become the accountable signatory.

Harvey will use the Series D to make its platform more capable and easier to govern. Its execution test is whether customers can move beyond isolated demonstrations while retaining the controls that professional work demands.

About Harvey

Harvey develops generative-AI products for legal and professional-services organisations. Its platform supports research, analysis, drafting and knowledge workflows within customer-controlled environments.

Company contact
Harvey
Media and company inquiries
Website: https://www.harvey.ai/
Announcement URL: https://www.harvey.ai/blog/our-series-d-building-the-future-of-professional-services