Our Take
Crosby's bet is structural, not technical: they've chosen a business model where AI handles volume, lawyers handle judgment, and fixed fees force efficiency instead of hour-padding.
Why it matters
The legal market is watching whether an AI-first law firm can sustain profitability while undercutting Big Law on cost and speed. Crosby's willingness to absorb losses on individual matters signals confidence in the model's long-term viability—or desperation to prove it works.
Do this week
GCs and procurement teams: request fixed-fee contract review pilots from MSO-model law firms and measure turnaround time and cost per agreement against your current Big Law baseline before 2027 budget lock.
Crosby operates as a two-entity structure: a licensed law firm paired with a for-profit technology and services company
Ryan Daniels and co-founder John built Crosby on an MSO (Managed Services Organization) model pioneered by Atrium, which attempted the same experiment in 2018 but ran ahead of AI capability. Daniels spoke to Atrium's founders before starting Crosby and confirmed they had the right vision but were too early.
The law firm entity hires experienced lawyers and serves clients. The separate corporate entity licenses technology and provides non-legal services like invoicing and marketing. This structure solves what Daniels calls the "incentive problem": traditional law firms bill by the hour, which discourages offloading work to AI. Crosby inverts that. The better the AI performs, the more profitable the model becomes.
Crosby started with high-volume contract review—NDAs, MSAs, DPAs, partnership agreements—because the liability profile is lower than negotiation-heavy work and the market actively seeks alternatives to Big Law pricing. Daniels acknowledges they sometimes lose money on individual matters to prove the model scales and to build client relationships.
The fixed-fee bet hinges on AI doing paralegal-grade work while lawyers retain final judgment
Daniels positions AI not as a replacement for lawyer judgment but as an accelerant. Contract review, he explains, is "extremely judgment based"—it's a negotiation between two parties, trading concessions. Good lawyers understand client playbooks and risk tolerance. AI can surface issues fast. Lawyers decide what matters.
This is not a commoditized, cookie-cutter operation. Crosby tailors AI outputs to each client's documented preferences and negotiation patterns. The output should match what a major law firm would produce. The difference is speed and cost.
If this model works, it redefines the competitive wedge against Big Law: not "we use AI to cut corners," but "we use AI to eliminate billing friction and pass the savings to you." Big Law firms, accustomed to hourly billing as the default, would need to fundamentally redesign their incentive structure to compete—a shift many are reluctant to make.
The risk is binary. Either Crosby's AI-lawyer pairing produces work that clients accept as equivalent to Big Law quality (and the fixed-fee model holds margin as volumes grow), or it doesn't. There is no middle ground in legal services: a contract is either defensible or it exposes the client to liability.
Evaluate MSO law firms on output consistency and turnaround, not brand pedigree
If you are responsible for legal spend and currently route contract work to Big Law firms, pilot a fixed-fee agreement with an AI-first law firm on a representative batch of agreements before committing multi-year volume. Measure: turnaround time from submission to final output, number of revision rounds, and cost per agreement.
Ask about the firm's approach to client-specific personalization. A firm that treats every NDA the same way is not solving the problem Crosby claims to solve. One that maintains a documented "playbook" for your company's risk tolerance and negotiation patterns is worth testing.
Be explicit about liability expectations. Fixed-fee does not mean "we assume your legal risk." Understand what coverage the firm carries and what remains your responsibility as client. The cost savings only matter if the legal exposure is equivalent.