Summary: Crosby is building a new type of law firm that focuses exclusively on "contract negotiation" using AI. Instead of the traditional billable-hour model, they use per-document pricing, close collaboration between AI and lawyers, and rapid contract processing to fundamentally transform the existing legal services paradigm. This video vividly covers structural choices, the reality of AI adoption, culture, and future outlook.

1. Crosby's Origin Story and Innovative Model

Crosby is not simply a legal software startup. Rather, it has embedded AI engineers alongside lawyers within a law firm structure to automate the entire contract negotiation process. The key innovation is building an environment that enables rapid experimentation and improvement — startup-style — overcoming the drawbacks of traditional partnership structures, such as limited technology investment and inflexibility for long-term innovation.

"Crosby is an AI-first law firm. We focus exclusively on contracts. Our goal is to automate agreements between people."

This structural experiment translates directly into rapid feedback and workflow improvement. Lawyers and developers are even physically seated in alternating desks so that "real-time collaboration and instant feedback" occur naturally. The result goes beyond mere software development to enable service innovation grounded in real-world practice.

2. How AI and Lawyers Collaborate: What Real Automation Looks Like

The core innovation lies in the careful design of what work humans do and what AI handles. The Crosby team has complex contract review and risk assessment still handled by human lawyers, while repetitive recording, organizing, summarizing, and explaining are handled smoothly by AI.

"The magical experience of delegating the work you never wanted to do to AI — we witness this on a weekly basis!"

In particular, they focus on AI's ability to explain "why this clause needs to be changed and why this word should be included," achieving the added value of reducing negotiation rounds and persuading counterparties.

For each core contract document (NDA, MSA, DPA, etc.), AI plays an optimized role:

• AI paralegal agents handle initial distribution and classification
• Based on complexity, human lawyers or more sophisticated AI then intervene

This creates a seamless "human-machine-human" coexistence.

3. Per-Document Pricing Instead of Billable Hours — A Groundbreaking Business Model

Eliminating the legal industry's signature practice of "hourly billing" is another major strength. Rather than "how many hours were spent," they simplify pricing to how much per contract document, giving clients both predictable costs and speed simultaneously.

"The billing model innovation was easier to decide than you'd think. It almost automatically became 'we don't charge by the hour.'"

For this "per-document pricing" model to work, AI must predict in advance how many rounds a contract will go through and how long it will take, which is technically quite challenging. But this is precisely what makes it Crosby's differentiator and barrier to entry.

"The real area where we innovate is that we can predict almost exactly how long a job will take before it starts and quote a price. That's harder than you might think!"

4. The Power of Crosby's Data and Feedback Loops

What matters in Crosby's approach is the data accumulated from service delivery and the systematic feedback loops that track and improve operations.

• Every aspect of the work process is meticulously measured — how many human interventions occurred, how long things took
• This continuously distinguishes what AI and humans should handle, and what can be fully automated, with repeated validation

Additionally, since individual standards for "good contracts" vary and industry benchmarks are fluid:

• Per-customer model fine-tuning
• Hyper-granular feedback to push accuracy to 99% — this directly translates into competitive service quality

"Getting to 90% is easy. But the real challenge is pushing to 99%, 99.99%."

5. The Real Customer Value Is 'Deal Velocity'

The greatest strength that Crosby's clients experience is "overwhelming speed of contract processing." For startups, everything — product launches, sales, hiring — moves at lightning speed, so contracts stalling for a day or two is a major burden. Leveraging AI, Crosby has dramatically shortened this time, achieving an average of under 1 hour.

"Contract negotiation is the API connecting businesses to customers. This hasn't changed in over 40 years. Our goal is to take it from one week to one hour, and eventually to minutes."

What used to take "5–6 rounds" has been reduced to "2–3 rounds" thanks to AI's accurate recommendations and explanations, significantly shortening the entire process.

Furthermore, "quality" must be maintained above a certain threshold alongside "speed":

• Customized results aligned with risk tolerance and business needs
• A system that "escalates" only to the points where lawyers must intervene

Both sides are harmonized through this approach.

6. Tech Stack, Models, and Operating Culture: Why Only Crosby Can Do This

1) Real Data Built Internally

While large public datasets (SEC Edgar, etc.) are widely used, actual "real contract negotiation data" generated in practice is never publicly released. This is why Crosby is building its own best-in-class "field data," which becomes the foundation of long-term competitive advantage.

2) AI Researchers and Lawyers, Side by Side

"We literally arrange desks in lawyer-engineer-lawyer order and create a structure where the feedback cycle runs infinitely. This is the biggest difference from traditional law firms and AI startups."

3) Localized, Personalized Workflows

At each level — paralegal, junior associate, senior — separate AI agents are designed with each company/team's unique know-how and requirements precisely reflected.

4) Service Usability and Internal Culture

They create a user experience that blends seamlessly into everyday collaboration tools like Slack, so both lawyers and clients can move freely.

In practice:

• Lawyers are actively encouraged to design and create their own "prompts"
• AI is instructed to explain "why this particular word should be used" They continuously promote self-directed learning, growth, and innovation.

"We try to actively praise and reward lawyers who write their own prompts and improve workflows."

7. Building the 'Next Generation Law Firm' in New York: Crosby's Culture and Talent

The backstory behind building an AI-based law firm in New York is also fascinating.

• Leveraging New York's unique strengths: existing startup ecosystem, concentration of financial talent, creative industry density
• Connecting diverse talent — "experienced founders," "ambitious new hires" — in "domain expert + AI" configurations

"New York is both a laboratory for ideas and a site for execution. This year, companies like Ramp have become a kind of 'founder factory.' We think the next generation of founders will pour out within 4–5 years."

The organizational culture is actively flexible, emphasizing:

• Public recognition, process map sharing, and encouraging innovative ideas within repetitive work

8. Fully Automated Legal Services and the Future of the Legal Market

Regarding the scope of legal services that could be fully handled by AI without lawyers, Crosby's founders emphasize "creating new markets."

• Top-tier law firms (the upper 8%) will remain stable for a while
• The remaining 92% of small firms and individual services (leases, child support, etc.) were markets where "practically no one was helping," so AI's full automation would actually greatly resolve 'legal blind spots'

"Where there are no alternatives, no one loses anything when AI handles everything. It actually opens a new market."

They envision a future where AI enables a single lawyer to handle 500 cases simultaneously, while core work (contracts, litigation) remains under lawyer oversight.

9. Key Metrics and AI's Limitations: The Quality-Speed-Trust Tradeoff

Crosby's top priority metric is Total Turnaround Time (TTA) — the total time from contract receipt through all negotiations to final completion.

"Traditional law firms earn more the longer this takes. We're the opposite — reducing this number is our growth engine."

Simultaneously, finding the ideal balance of "automation + lawyer intervention" to match each client's risk tolerance and quality standards is emphasized as paramount.

They also share the "HURT (Human Review Time)" metric:

• Less "HURT" (human intervention), higher quality is the ongoing pursuit.

"Contract negotiation is an abstraction of the human-to-human agreement process. No matter how advanced AI becomes, this fundamental element won't be easily replaced."

10. The Ultimate Future: Agent vs. Agent and the Changing Role of Legal Professionals

The most intriguing future scenario:

• Both sides (e.g., buyer and seller) would each have their own AI agents that actually negotiate with each other
• Customized agents pre-loaded with each side's data separation, risk appetite, negotiation limits, and all acceptable terms would simulate in real time for maximum efficiency

"Each participant would have different risk tolerances, bottom lines, and available negotiation time. AI agents with those characteristics built in would lead transactions — it would be a beautiful world."

Ultimately, in the future, the role of senior partners (lawyers) shifts toward managing armies of agents:

• Instead of large legal teams, hyper-specialized AI-first companies
• New legal demand that didn't exist before (eliminating blind spots, ultra-fast service)

The very structure of the legal market will change.

11. Advice for Law Students and Future Legal Professionals

In the flow of AI development and legal service innovation, future legal professionals should:

• Rather than fixating on existing conventions (footnote formatting, form compliance, etc.)
• Emphasize "questioning everything and personally experimenting with new tools and approaches"

"Question everything. Even the way your professor teaches footnotes — is it really necessary? There's too much dogma. And accepting old methods uncritically is dangerous."

Additionally:

• Understanding the flow of automation while maintaining the wisdom and balance of legal academia
• Gradual innovation starting from small changes — preview, apprentice-style learning, AI prompt utilization is cited again as the key to future competitiveness.

Conclusion

Crosby's case is a vivid laboratory demonstrating how quickly and fundamentally AI can transform law firms and the legal services market. The core is not about taking away jobs but creating new markets by providing services that were previously unavailable, through organic collaboration between lawyers and engineers. They are building a future where human expertise is amplified while maintaining "speed + quality + trust." If you're interested in legal innovation, AI and automation, or organizational design, the Crosby model is a noteworthy role model.