Skope Is the AI Associate Your Law Firm Can Actually Afford

The Macro: Small Law Firms Are Drowning in the Wrong Work

There are roughly 450,000 law firms in the United States. The vast majority of them are small. Solo practitioners and firms with fewer than 20 attorneys make up over 85% of the market. These firms handle real cases for real people: personal injury, family law, immigration, criminal defense, small business disputes. They are busy, understaffed, and spending an absurd percentage of their time on work that does not directly serve their clients.

I am talking about legal research that takes six hours when it should take one. Contract review that requires reading 40 pages to find three clauses that actually matter. Document drafting that starts from a template but requires enough customization that it might as well have started from scratch. These tasks are essential but they are not the high-value work. The high-value work is strategy, client counseling, and courtroom advocacy. Everything else is overhead.

The legal tech market is projected to reach over $35 billion by 2027. That number is inflated by enterprise products built for Am Law 100 firms that will never sell to a three-person practice in Phoenix. The tools that big firms use, like Relativity for e-discovery or Kira for due diligence, cost six figures annually and require dedicated IT staff to maintain. They are irrelevant to the firms that need help the most.

Clio dominates practice management for small firms. LawDroid handles chatbot intake. CaseText (now part of Thomson Reuters) was the early mover on AI-powered legal research. Harvey is building for big law with big funding. But the specific combination of research, drafting, and document review in a package that a small firm can actually adopt without an IT department is still an open problem.

The firms that need AI the most are the firms least equipped to evaluate, purchase, and implement it. That is the gap.

The Micro: An AI That Edits Your Word Documents With Tracked Changes

Ben Smith and Connor Park started Skope in San Francisco as part of Y Combinator’s Summer 2025 batch. Smith is the CEO, Park is the CTO. The team is two people.

Skope has two core features and both of them are smart product decisions.

The first is an AI assistant for legal work. You can research case law, draft documents, and review contracts through a conversational interface. It generates Word documents from templates, reviews contracts and flags issues, and researches case law with cited sources. The citations are the important part. An AI that produces legal research without citations is useless to a lawyer, because lawyers need to verify everything they submit to a court. Hallucinated case citations have already embarrassed several attorneys publicly, and nobody in the legal profession has forgotten those stories.

The second feature is a Microsoft Word add-in. This is where I think Skope makes its smartest bet. Lawyers live in Word. They draft in Word. They redline in Word. They send Word documents to opposing counsel. The entire legal profession runs on tracked changes in Microsoft Word. By building an AI that operates inside Word using tracked changes, Skope meets lawyers exactly where they work.

The interaction model is clean. The AI suggests edits using the same tracked changes interface that lawyers already use every day. The attorney can accept or reject each revision without leaving the document. This is not a new workflow. It is the existing workflow with AI inserted into it. That matters enormously for adoption. Lawyers are conservative technology adopters by professional necessity. A tool that requires learning a new interface will not get adopted. A tool that works inside the interface they already use has a real shot.

Skope is SOC 2 Type II certified, which is unusual for a company this early. They advertise zero data retention for AI training, and enterprise-grade encryption. For law firms handling sensitive client information, these are not optional features. They are prerequisites. The fact that Skope invested in security certification before scaling suggests the founders understand their buyer.

The product is apparently quick to deploy. “Most law firms are up and running in minutes” is the claim. If true, that eliminates one of the biggest barriers to legal tech adoption. Demo calls, onboarding sessions, and training weeks kill deals with small firms because small firms do not have time for any of that.

Pricing is not public, which is standard for legal tech products that likely adjust by firm size.

The Verdict

Skope is doing two things right that most legal AI startups get wrong. First, it works inside Word instead of asking lawyers to use a separate application. Second, it secured SOC 2 certification early instead of treating security as a growth-stage problem. Both decisions suggest a team that understands how law firms actually buy and use software.

What I want to see at 30 days is whether the case law research produces reliable citations. This is the make-or-break feature for any legal AI tool. One hallucinated case and the product is dead in that firm forever. At 60 days, I want to know if the Word add-in is stable across different versions of Word on Windows and Mac, because Microsoft Word compatibility is a minefield that has killed better products. At 90 days, the question is whether small firms are actually buying or just trying the free tier and leaving.

Harvey is raising hundreds of millions to build AI for big law. Skope is building for everyone else. I think the “everyone else” market is bigger, harder to reach, and more rewarding if you get the product right. The Word integration is the kind of unsexy, obvious-in-hindsight decision that wins markets. Most AI startups want you to use their shiny new interface. Skope just wants to be useful inside the interface you already have.

That is a better strategy than it sounds.