AI Does the Contract Review – Real Lawyers Sign Off in Under an Hour

01 /The Concept

Crosby is an "agentic" law firm that helps client companies conduct fast, structured reviews of contracts – whether those contracts arrived from counterparties or were drafted internally.

"Fast" means genuinely fast: contract review is completed in under an hour. "Agentic" refers to the fact that the bulk of the work is done by an AI agent built by Crosby.

Although the AI agent handles most of the heavy lifting, the key differentiator is that the final output is delivered by real lawyers – who review, correct, and improve what the AI produced.

This doesn't slow anything down. Crosby promises 24/7 availability and no delays.

Pricing is determined upfront at the moment the contract is received, based on length and complexity. No more billing by the hour – which is the norm at virtually every other law firm.

The process is three steps:

- The client submits a document to Crosby. They can even configure automatic forwarding of all incoming counterparty contracts as they arrive.

- Crosby's AI agent analyzes the document, generates a plain-language summary, and produces a list of recommended edits and additions.

- A human expert reviews the AI's output for accuracy and checks for hidden issues or risks the AI might have missed. Where necessary, they make corrections and additions, then send the final result to the client.

One important distinction: Crosby's AI doesn't just analyze the legal language. It also evaluates the "market" dimension of the contract – whether the proposed terms reflect current market norms for that type of deal. Effectively, it helps clients negotiate better terms from a position of knowledge.

The AI draws on market intelligence from contracts it has previously reviewed, publicly available agreements, and closed legal databases.

Another edge: the AI has memory. With each new contract from a client, it accumulates knowledge about that client's business and context – which meaningfully improves the analysis of every subsequent document.

Crosby is still in closed beta but has already raised a $5.8M first round led by Sequoia Capital.

02 /Why It Matters

Crosby describes itself as the world's first "hybrid" law firm – meaning the output clients receive emerges from close collaboration between AI and human lawyers, not from either working alone.

It's worth noting where humans sit in that collaboration: after the AI. They refine what the AI produces rather than directing it upfront. Many still assume the most important human role is before the AI – crafting precise prompts so the AI can generate output good enough to use as-is. That gave rise to talk of the "prompt engineer" as an emerging profession.

But Crosby's model suggests the "output engineer" – the specialist who improves what the AI generates – may be equally valuable, or more so.

The general architecture of a hybrid firm runs like this: AI handles the core analytical work; humans on both ends shape and polish the result. One group engages with the client to understand their requirements and frame the task clearly; another reviews and improves what the AI delivers before sending it out.

Valid ([related review](/review/prodajot-ne-nachinka-a-upakovka)) runs exactly this model – a hybrid AI ad agency that raised $5.5M in February. Human specialists take the client brief, translate it into a clear task for the AI, the AI produces creative drafts, and human specialists refine and deliver the final output.

GrowthX ([related review](/review/novaja-biznes-model-dlja-bystrogo-i-pribylnogo-rosta)) applies the same approach to SEO and raised $12M in its first round in May.

The GrowthX story is instructive. The founder originally built the AI-powered SEO system for his own employer. Then he started explaining to other companies how they could implement the same approach internally – which required adopting new AI tools, hiring staff to write briefs and edit AI output, and then managing the resulting operation. Companies, it turned out, didn't want to do any of that. They just wanted the result.

So he left his employer, built an agency running the same model, and started delivering that result directly to clients. Monthly revenue climbed from $0 in July 2024 to over $600K in April of this year.

03 /Opportunities

In the real world, you can't count on AI producing a perfect result at the press of a button. AI needs to be wrapped by specialists – people who prepare clear briefs and task definitions going in, and people who improve the AI's output coming out.

The vast majority of companies aren't going to implement those hybrid workflows internally. It takes additional effort, restructuring, and new hires. Even though those workflows consistently produce better results than any off-the-shelf AI platform used as a "magic button"

But companies still want better results. And they're willing to pay – as long as the deal is "just money, no headache" That opens a large and genuinely new market for hybrid agencies capable of delivering expert-quality outputs across a wide range of service categories.

The startup opportunity here is specific: build a hybrid agency in a domain you understand well, where the AI-plus-expert model produces measurably better results than either can alone. Crosby did it in legal contract review. GrowthX did it in SEO. The model is proven; the domains that haven't yet been claimed are the ones worth targeting – and there are still many of them.