Wedge

The initial wedge targets high-growth B2B SaaS legal departments drafting routine Master Services Agreements and Data Processing Agreements. This niche provides high-velocity, highly standardized contracts where drafting speed directly accelerates sales cycles, making the return on investment immediately measurable. From this beachhead, the product expands horizontally into adjacent corporate departments for employment or procurement contracts before moving upmarket to handle bespoke drafting for specialized law firms.

Timing

Recent advancements in long-context LLMs allow systems to ingest hundreds of pages of precedent contracts and output highly accurate, formatted legal text without hallucinating clauses. Buyers now trust generative models to produce first-pass legal drafts, shifting their behavior from outright skepticism to demanding drafting acceleration.

Why This ICP

Mid-sized law firms and corporate in-house teams face acute margin pressure and a high volume of routine transactional work like NDAs, MSAs, and DPA agreements. They possess the transaction volume to justify the software but lack the budget of Big Law to hire armies of junior associates for manual assembly.

Size Of Prize

There are approximately 50,000 mid-to-large corporate legal departments and law firms in the US and UK. At an average annual spend of $15,000 per entity on drafting labor displacement and template maintenance, the addressable market is roughly $750M.

Gap Narrative

Corporate legal departments and mid-sized law firms spend hundreds of hours manually assembling standard agreements from fragmented clause libraries and past templates. Current document automation tools require hardcoded conditional logic that breaks when drafting complex, negotiated variations. This opportunity provides an engine that generates context-aware, fully formatted first drafts based on plain-language prompts and precedent analysis.

Defensibility

Defensibility compounds through custom precedent libraries and workflow lock-in. As the engine ingests a specific firm's historical redlines and preferred fallback clauses, the drafting output becomes tightly customized to their specific risk tolerance, creating a steep switching cost. Competing horizontal text generators remain generic, while this system becomes deeply embedded in the firm's unique institutional knowledge.

Why This Thesis

An Agent thesis fits perfectly because legal drafting requires iterative generation, retrieval of specific precedent, and formatting adjustments. The agent acts as a junior associate, taking high-level instructions to retrieve clauses and synthesize drafts, which aligns directly with the senior lawyer's existing workflow of reviewing subordinate work.

Hardest Part

Maintaining perfect logical consistency across deeply nested cross-references and defined terms in a 50-page document without introducing fatal liability risks or hallucinations.

Min Viable Scope

Automate first drafts of NDAs and Master Services Agreements for mid-market software companies. Leave out complex M&A, real estate, bespoke joint ventures, and automated negotiation or redlining capabilities.

Cold Start Problem

The engine needs thousands of high-quality negotiated templates to learn acceptable deviation ranges. Break this by ingesting publicly filed EDGAR contracts and partnering with a boutique firm to curate a seed library of gold-standard playbooks.

Time To First Value

1-2 weeks of playbook mapping and template ingestion

Data Moat Available

true

Technical Difficulty

High

SAM

~$800M-1.2B US mid-to-large law firms

SOM

~$20-50M

TAM

~50k transactional law firms globally x ~$50k/yr = ~$2.5B

Growth Rate

~12-18%/yr, driven by associate salary inflation and client pushback on billable hours for routine drafting

Paid Comparable Spend

~$150k-200k/yr per junior associate for manual drafting labor, plus ~$10k-20k/yr on legacy document assembly software

Kill Thresholds

• Edit distance to final draft exceeds 40% after 30 days of usage
• Fewer than 2 exported contracts per active user per week
• Trial-to-paid conversion falls below 15% at the $2k price point
• System-generated hallucination rate triggers manual rewrite in >10% of drafts

Leading Metrics

• time-to-first-draft-generation
• edit-distance-from-engine-draft-to-final-version
• average-drafting-sessions-per-user-per-week
• trial-seat-expansion-velocity

What Proves Right

Transactional lawyers generate first-draft contracts from prompt to output in under 5 minutes without opening legacy templates. Weekly active users run at least 4 drafting sessions per week with the engine-generated text requiring under 20% modification to reach the final version. Mid-sized firms convert to paid subscriptions at $2,000 per seat annually within 14 days of initial testing.

What Proves Wrong

Associates generate initial drafts but discard them because hallucinated clauses create more manual review work than starting from an existing firm template. Trial users abandon the application after 3 days when the engine fails to adopt the specific styling and defining language of their practice group. Partners refuse to approve the output, forcing teams back to manual document assembly out of liability concerns.