Harvey: Legal AI's frontrunner – or Big Law's power-up?

Harvey is one of the fastest-growing names in legal tech. Backed by OpenAI and Sequoia, it claims to offer end-to-end support for law firms using AI: drafting, research, contract analysis, redlining, you name it.

And it’s clearly doing something right. Since 2022, it’s secured $300 million in funding, partnered with LexisNexis, and signed up nearly a third of the Am Law 100. In benchmark tests, it’s outperformed other models on a range of legal tasks, including data extraction, summarisation and clause comparison.

But for all the buzz, Harvey is not a tool most of us can use. It’s designed and priced for Big Law and enterprise firms. So what does that mean for everyone else?

Harvey works best when the stakes (and budgets) are high

Harvey isn’t a plug-and-play AI chatbot. It’s a sophisticated, modular platform designed to work across an entire firm’s legal operations. Tools like Harvey Assistant and Harvey Vault are trained on firm-specific workflows and can extract key terms from thousands of contracts with high precision. It integrates with Microsoft Word, SharePoint and Icertis, and it’s built to scale across enterprise infrastructure.

That’s great if you’re a multinational firm with a dedicated innovation team and a huge caseload. But it’s less relevant if you’re a small legal business.

For firms like Plume, Harvey isn’t an option. It’s invite-only, enterprise-tier, and currently priced and structured in a way that puts it out of reach for most challenger firms.

The transparency gap: Is AI helping, or hiding?

One of Harvey’s biggest selling points is how deeply it integrates into law firm workflows. But that also creates a transparency challenge. If your panel firm uses Harvey to analyse a contract or draft an advice note, how do you know? Are you being billed for human hours when the first draft was machine-generated? Was the AI output reviewed properly? And is the use of Harvey improving the outcome, or just speeding it up?

These are fair questions, but the answers aren’t always forthcoming. Harvey’s client firms aren’t required to disclose when or how they use it. That puts the onus on clients to ask questions about quality assurance, data privacy, and fair pricing.

Is Harvey worth the hype?

If you’re a global law firm, possibly. Harvey’s lawyer-trained models, high accuracy scores and deep integrations make it a compelling proposition for high-volume, high-complexity work. And its development roadmap – including agentic AI that plans and completes multi-stage legal tasks – points to an ambitious future.

But that future isn’t evenly distributed. For now, Harvey is reinforcing existing hierarchies in legal service delivery: boosting the efficiency and profitability of the firms already at the top.

For the rest of us, the legal AI landscape is still fragmented. There are powerful tools out there – some generalist, some domain-specific – but nothing as vertically integrated or elite-focused as Harvey.

Final thought

Harvey is redefining what’s possible in legal AI. But it’s also redefining who gets access.

As regulators and clients start asking harder questions about transparency, value and accountability, the firms using Harvey may need to shift gears. And for the wider legal sector, Harvey is a wake-up call: the AI future is coming – but we need to shape it, not just watch it happen.

the plume press

THE NEWSLETTER FOR IN-THE-KNOW IN-HOUSE LAWYERS

Get the lowdown on legal news, regulatory changes and top tips – all in our newsletter made especially for in-house lawyers.

sign up today