Crosby is an AI-first law firm specializing in contract negotiation automation. Instead of building legal software, they built an actual law firm where lawyers and AI engineers collaborate to automate the human negotiation process. They abolished billable hours in favor of per-document pricing and are achieving sub-one-hour contract processing times. In this video, Crosby's founders explain in detail how the law firm structure accelerates their innovation cycle, how they use AI to predict negotiation outcomes, and their vision for AI agents that can simulate entire contract negotiations between parties.

1. What Is Crosby?

The video introduces Crosby, co-founded by Ryan Daniels and John Sarihan, as an AI-first law firm. Crosby focuses exclusively on contracts, aiming to automate the human negotiation process. They explain that contracts exist everywhere in our lives — leases, employment agreements, most business transactions — and the best way to automate this complex process is to build a law firm internally, understand how lawyers work, and replace their processes with AI to accelerate deal closure.

Crosby discusses two main reasons for choosing a law firm over a legal software company:

• Structural perspective: The legal services market has historically depended heavily on human capital, but technological advances — especially in the last 3 years — demand new innovation models that can replace human work with technology. Traditional law firm partnership structures were limited in investing in speculative activities like technology development, and Crosby believes it can solve these structural problems to create faster innovation cycles.
• Practical perspective: Having domain expert lawyers working side by side with engineers creates a unique feedback loop. This goes beyond simple evaluations — it provides an environment where they use and experience the actual product, identify important workflows and the biggest gaps through user research, and improve contract review speed and even individual clause accuracy.

"We literally seat lawyers, engineers, lawyers, engineers in alternating desks to ensure feedback cycles and encourage them to truly collaborate."

2. Differentiation from Previous Attempts and How AI-Lawyer Collaboration Works

Crosby's founders mention that previous attempts to combine software and services fell short for two main reasons:

• Lack of close collaboration between lawyers and engineers: Crosby builds real-time feedback loops through a structure where lawyers and engineers work side by side.
• AI technology limitations: In the past, it was difficult to delegate much of the qualitative, thoughtful work lawyers perform to machines. Now, the latest AI technology makes it possible to automate these complex tasks.

Crosby emphasizes two main outcomes of this collaboration:

1. Product Velocity: When contracts come in, they are processed quickly to meet SLAs, identifying where human intervention is needed and targeting automation opportunities.
2. Instrumentation: Various metrics are measured and analyzed — contract turnaround time, number of human interventions — to maximize efficiency. This goes beyond just shipping product to verifying contribution to end goals like customer satisfaction.

"Ultimately, clients trust us for high-quality contract review. Fast-growing companies like Cursor and Clay are growing faster than ever, and we need to meet their standards of excellence and speed."

Crosby also explains that legal services are credence goods requiring expert review. Lawyers guarantee quality and assume liability, maximizing AI's potential while providing safe, trustworthy service to clients.

3. Pricing Innovation: Per-Document Pricing Instead of Billable Hours

Crosby boldly abandoned the traditional billable hour model for per-document pricing. This was a strategic decision to align the company's incentives with client goals.

"People have predicted the death of the billable hour for 70 years, but it's actually a relatively new concept that gained popularity in the 1950s. It's remarkably durable and quite reasonable for complex, unpredictable work."

Crosby sees the ability to predict how long a piece of contract work will take before it begins as the core of their innovation. This is possible by accurately predicting how many revision rounds a contract will go through and how much review it will require. This enables transparent, predictable pricing for clients while aligning incentives with the value Crosby delivers.

4. The Actual Process of Contract Automation and AI's Role

Ryan explains how lawyers actually work in contract negotiations and what types of contracts Crosby currently focuses on.

• Current focus: Crosby concentrates on contracts needed for B2B product sales — NDAs (Non-Disclosure Agreements), MSAs (Master Service Agreements), and DPAs (Data Processing Agreements). NDAs are relatively simple, but MSAs and DPAs can be 15 pages long and over 80 times more complex than NDAs. The ultimate goal is to handle contracts over 1,000 times more complex, like merger agreements.
• The lawyer's role: Lawyers review all contract terms, using mental models to negotiate conditions that are safe and not risky for the business. They protect companies from excessive risk exposure and understand market-standard terms based on industry guidelines and benchmarks. However, these standards exist as abstract concepts in lawyers' minds, and even two lawyers at the same firm may disagree.
• AI's role: Crosby uses AI to convert lawyers' "intuition" and "experience" into quantitative probabilities. For example, predicting how likely a specific governing law jurisdiction will be accepted in a particular negotiation. This enables presenting statistically grounded risky or safe choices rather than just guessing at a "reasonable contract."

"It's about converting the vibes and rules of thumb in lawyers' heads into actual probabilistic, quantitative numbers — 'How much should we accept Delaware, California, or New York governing law in this particular negotiation?'"

AI Tech Stack and Workflow

John explains the technical side of Crosby:

• LLM utilization: Crosby uses foundation models from partners like OpenAI, Anthropic, and Google. However, since these models are not specialized in contract data, they build their own data and fine-tune models. Company-specific contract data that foundation model companies cannot access becomes a valuable asset.
• Context Engineering: Initially focused on automating manual workflows by providing AI with the tools and information lawyers need. Then testing and adjusting how accurately the language model can replicate human lawyer behavior.
• Individualized models: Crosby focuses on tuning language models to match individual lawyers' consistent judgments. Since two lawyers at the same firm may disagree on specific contracts, individualized fine-tuning aims for 99.99% accuracy.

"I think it's really important to tune to a specific person when you really need to. As Ryan said, even two lawyers at the same firm can have disagreements on a particular contract. If you tune more accurately to one individual, you can say 'this is right and that is wrong' — because that person is internally consistent."

5. Customer Value Proposition and Future Vision

Crosby's customer-facing value proposition is 'deal velocity.' Fast-growing startups need everything to move quickly, so the contract negotiation process must also be accelerated.

• Speed improvements:
  • Reduced turnaround time: Shortening the time from when a client sends a contract to when they get it back (currently median under 1 hour), with the ultimate goal of reducing this to minutes.
  • Fewer negotiation rounds: AI predicts which clauses to agree on and which to dispute, reducing unnecessary back-and-forth and shortening overall negotiation time.
• Quality and trust: Lawyers are included in the system to maximize AI's potential while ensuring quality and safety through lawyer oversight. Crosby carries malpractice insurance for all legal work they provide, calling it a core business element.

"Right now our median turnaround time is under one hour. Can we get that down to minutes? While simultaneously ensuring the right terms get to the right lawyer at the right time."

AI Agent Roles

Crosby describes the role of AI agents as follows:

• Summarization and explanation: AI excels at predicting comments needed to explain contract changes. Thoughtful explanations of contract changes help counterparties understand and reduce negotiation rounds.
• Information provision: AI provides lawyers with timely, relevant information about the client's business (e.g., how Cursor IDE works, how Clay uses sales data) to speed up their work.
• Agent layering: Currently, Crosby has built an agent serving as a paralegal that triages all incoming work from clients and assigns it to appropriate lawyers. The next step is developing agents for other roles — junior associate, senior associate, junior partner — to replace specialized tasks.

Ultimately, Crosby envisions a future where AI agents reflecting each party's preferences simulate negotiations and provide auditable records of the process. This would essentially become a collaboration platform for reaching agreements faster in cooperative negotiations.

6. Building a Business in New York with a Unique Culture

Crosby notes that building an AI company in New York differentiates them from other locations (e.g., San Francisco).

• Talent pool: New York has long produced seasoned engineers from AdTech, FinTech, and financial trading firms (Jane Street, HRT, Citadel, etc.). Over the past 5–10 years, many senior engineers from these backgrounds have flowed into startups, contributing to successful growth trajectories. Companies like Ramp serve as "founder academies," spreading startup culture across New York.
• Domain expertise: New York possesses deep domain expertise across finance, creative industries, healthcare, and law. This is a major advantage for applying AI technology to real industries and solving specific problems.
• Culture: Crosby describes New York's culture as having a "bias to starting things." They tell employees "this will be a 4–5 year stepping stone" while emphasizing autonomy and ownership. Lawyers are encouraged not just to fill hours but to actively contribute ideas for improving processes and reducing total review time.

"We do a lot of repetitive, same-kind-of-work, but there are ways to step out of that and think bigger. So we actually incentivize lawyers to keep pushing total review time down."

7. The Future of the Legal Industry and AI's Role

Crosby predicts the legal industry will undergo major changes over the next decade.

• Growth of in-house legal teams: Over the past decade, in-house teams have grown much faster than law firms. As legal work becomes increasingly complex, companies will need larger in-house teams. These teams have fewer constraints than law firms, hiring paralegals, legal operations specialists, and other diverse personnel to perform work in creative ways.
• AI-first specialty firms: AI-first companies specializing in legal services will emerge to fundamentally change how legal services are delivered. Crosby sees this as a small part of a larger transformation — the beginning of a true golden age for legal market innovation.
• Changing lawyer roles: AI will automate paralegal and junior associate work, enabling senior lawyers to manage armies of AI agents and wield greater influence. For example, in personal injury firms, AI handles case intake, bill drafting, and review while lawyers focus on core work like court appearances.
• Legal work that will be automated: Personal legal services that are currently underserved (child support, lease agreements, etc.) are highly likely to be fully automated by AI. This isn't about taking lawyers' jobs — it's about AI handling work nobody was doing, creating new markets.

"I think this is a golden age for what lawyers will be able to do in the next 5–10 years. It will be unlocked by this."

Advice for Future Law Students

Crosby's founders advise current law students to question everything. Critically examine existing practices and traditions while understanding and balancing the excellence and importance of legal academia. Learning to leverage AI is emphasized as essential.

"Law students should question everything thoroughly. Even the way professors write footnotes. Is it really necessary? There's too much dogma. And you need to balance that with understanding the excellence of legal scholarship. That's what makes society function."

8. Key Metrics and Motivation

Crosby's Northstar metric is Total Turnaround Time (TTA) — the total time from contract receipt through all negotiations to final completion. This stands in stark contrast to traditional law firms that profit by extending this time. Crosby uses this metric to align incentives with clients, perform faster and better work, and minimize negotiation rounds.

"Total turnaround time means the time it takes from when a contract comes in to when it goes out. It's all the time spent on a contract including all back-and-forth. As Ryan said, a contract negotiation might have five round trips or two. The sum of all the time Crosby spent reviewing the contract is our key metric."

Beyond the target metric, Crosby maintains consistent quality through guardrail metrics. They have dedicated teams of lawyers and engineers managing AI agent environments and quality metrics, verifying that client requirements, risk profiles, and negotiation interests are met. They also use "HURT" (Human Review Time) as a metric, using fun, relatable metrics to motivate the entire team.

"We have another metric called 'HURT.' It stands for 'Human Review Time.' So we're trying to reduce the amount of 'HURT.'"

Conclusion

Crosby goes beyond simply applying AI technology to legal services — it innovates the fundamental structure and operation of law firms, placing "deal velocity" as its top priority. Through tight lawyer-AI engineer collaboration, the abolition of billable hours, and strategic AI agent deployment, it delivers unprecedented speed and quality in contract services. Built in New York's unique environment, Crosby's culture combines entrepreneurial spirit with domain expertise to usher in a new golden age for the legal industry. Long-term, they envision a future where AI increases legal service accessibility, providing high-quality legal services to the general public — a "Robin Hood" vision. Their story is a fascinating case study of how professional services can transform and innovate in the AI era.