A renewal pitch hit my inbox this week from one of the two big legal research incumbents. Their research platform has AI built in now — named frontier models, specialized legal workflows, drafting and analysis and document review, all “within the secure environment” of the platform. Nine months free if I sign before the end of the month.
I wrote back. I’d bet the salesperson is getting this reply more often lately, and the reasoning behind it is the whole ballgame for how a firm should buy legal technology right now. Here it is, lightly cleaned up:
I have zero interest in accessing an AI assistant inside your environment. You have no access to outside APIs or tools, so that is a nonstarter. I do, however, have a huge interest in allowing my AI environment to access you.
That’s the whole fight, in two sentences. Not which model. Not whose “skills.” The direction of the integration.
Your assistant is the office now
Every walled-garden AI pitch assumes the same thing: your work happens inside their product, so the assistant ought to live there too.
That made sense in 2023. It doesn’t now.
An AI assistant that runs in your environment — on your machine, under your control — isn’t a chat window. It sits on top of every tool you use. Mine reads the file system where my matters live. It queries my practice management system. It reads and labels the firm’s email. It pulls Minnesota statutes as raw HTML from the Revisor’s site, because summarizers paraphrase, and a paraphrase isn’t good enough when you’re quoting. It drafts in Word from the firm’s templates, checks its own citations against archival sources, and runs the same verification gate before anything leaves the building. Every one of those makes the others better, because they all live in one place and see the same picture.
Now look at the walled-garden version. The assistant inside a research platform sees exactly one thing: the research platform. It can’t see the matter file the research is for. It can’t see the draft the authority is going into. It can’t check the calendar, can’t read the demand letter you’re answering, can’t know that the client agreement caps the fee. It’s an assistant with one book open and amnesia about the rest of your practice. And every hour you spend in it is an hour spent teaching your workflow to live inside somebody else’s subscription.
This only runs one way. A tool can join your office. Your office can’t fit inside a tool.
The plumbing’s already built
The vendor doesn’t have to invent a thing. There’s an open standard — the Model Context Protocol — built for exactly this: letting an AI assistant search and pull from a service’s data in a controlled way you can audit. My practice management connector speaks it. This section has written about it before. Building it is weeks of engineering, not years.
And you don’t have to take my word that it works for legal research, because the nonprofit side of this business already did it. CourtListener, run by the Free Law Project, puts its whole corpus — case law, federal dockets via RECAP, oral arguments, judges — behind a documented REST API anybody can build against. My firm hooked it up the same afternoon-project way we hook up anything else, and the assistant uses it the way an associate uses a library: goes in with a question, comes back with an answer and a citation you can check.
That’s what I mean when I tell the salesperson firms are switching. A nonprofit’s database doesn’t match the incumbents shelf for shelf, and I’ll get to exactly where it comes up short. What firms are switching to is the open setup, because it’s the only one where the assistant can do its job. The research service that plugs into it gets used. The one that makes my assistant stop, hand off to a human, log into a browser, and retype the question doesn’t — and “has AI inside” doesn’t fix that, any more than a branch library hiring a librarian fixes the fact that the books can’t leave the building.
What we did without them
Here’s what that looks like on the ground. Over the past two days, this firm researched, drafted, and published thirty long-form articles on Minnesota law — family law, tort, business torts, consumer protection, expungement, construction payment. Every statutory quotation came from the Revisor’s raw HTML. Every case was read in full from the Caselaw Access Project’s archival scans or the Minnesota Judicial Branch’s own slip opinions, with pin cites resolved against star pagination, not guessed. We read recent session laws in struck-and-underscored form and caught three separate 2026 amendments that the codified statute pages didn’t show yet. Federal statutes came from the U.S. House’s official code site, federal regulations from the eCFR, Supreme Court quotes from the Court’s own prints. Then an independent verification pass re-derived all of it before anything published.
What we spent on commercial databases to do it: zero. For primary authority — what the statute says, what the court held, what the session law changed — the free and archival sources aren’t a compromise anymore. They’re the ground truth the commercial databases are built on. An assistant that reads them directly, at machine speed, in bulk, wipes out most of what the subscription was paying for.
Most. Not all.
What they’ve still got
What’s left is real, and it’s exactly what I’d pay for if they sold it right.
The one thing our all-primary-source pipeline can’t do is subsequent treatment. We can confirm an authority exists, confirm what it held, confirm the quote is verbatim. We can’t confirm it wasn’t reversed, superseded, or abrogated last Tuesday. There’s no free citator. A citator is hard to build: it’s editorial judgment applied across an entire corpus, every day, without a break. KeyCite and Shepard’s are the two incumbents’ crown jewels, and they earned it.
Add the secondary literature — treatises, practice guides, decades of editorial work piled on top of the law — and you’ve got the complete list of what a Westlaw or Lexis subscription still gives you in 2026 that nobody else does.
So the renewal question isn’t “do they have AI.” It’s this: is the citator-and-treatise sliver worth full price when it’s locked behind a screen my assistant can’t get to? Because that’s what’s actually on the table. The workflow in the pitch — a lawyer, in a browser, on their platform, chatting with their assistant — isn’t my workflow, and more and more, it isn’t anyone’s. My assistant hits the research service a hundred times in an afternoon, two seconds at a time, mid-draft, mid-verification, mid-gate. Sold per seat behind a login, the sliver isn’t worth it. Sold like a utility — metered, per call, open to the tools that actually do the work — it’d be an easy yes.
So here’s my call, stated plain enough that you can check it later:
For a small firm, dropping Westlaw and Lexis is the rational move — unless and until they sell low-cost API access. Not AI inside their walls. Not a copilot in their browser. A library card my assistant can carry: metered access to the citator and the editorial layer, at a price that reflects what it costs them to answer one more API call, which is fractions of a cent. The first incumbent to sell that wins the agent era, because every firm building what this firm has built will route through them without a second thought. Every year neither one sells it, the free corpus gets deeper, the checking tools get better, and the sliver gets thinner.
That word “secure”
The pitch’s biggest word is “secure” — the AI runs within the secure environment of the platform. Look closer and that turns the confidentiality question upside down.
A research query is about the least sensitive thing my firm sends out. “Minnesota conversion elements punitive damages” tells nobody anything a docket search wouldn’t. What is sensitive is everything the walled garden asks you to carry in: the client’s documents, for review; the draft, for analysis; the facts, for research memos. The pitch says your data is safe because the AI lives here. But the only reason my data would be in there at all is that their setup makes me bring it. My assistant sends the vendor nothing but queries. Theirs needs my client file to be any use.
“Secure environment” isn’t a security argument. It’s a data-gravity argument wearing one as a costume.
What I’d do when the renewal notice shows up
If you’re staring at your own renewal notice, here’s the order:
- Find out what you actually use. Pull the usage logs. In most small firms, the big majority of sessions is plain primary-authority lookup — a statute, a case, a quick elements check. All of that’s free today, at API speed.
- Set up the free layer first. CourtListener’s API for case law and dockets. Your state legislature’s site for statutes and session laws — raw HTML, not a summary. The Caselaw Access Project’s archives for the old reporters. Wire them to your assistant through MCP and live on it for a month while the subscription’s still running. The overlap tells you what you were really paying for.
- Split out the citator. What’s left is mostly treatment-checking. Put a straight price on that: a per-lawyer Fastcase-tier plan (a lot of state bars include one as a member benefit), or the smallest incumbent tier that includes the citator. Treat it as a checking step, not a research home.
- Then ask the salesperson one question: can my environment get at your service — not your AI, your data — through an API, and what does it cost? Whatever they say is your renewal decision. As of this week, from one of the two incumbents, I’ve gotten silence on the API and nine free months on the walled garden. That’s an answer too.
This is an access-to-justice story, same as everything else in this section. A solo can check the law like a firm with a research department today because people who believed public law should be publicly computable pried the statutes, the opinions, and the session laws into the open. And every dollar of subscription overhead a small firm sheds is a dollar that doesn’t have to come from a client who could barely afford the lawyer in the first place. The incumbents can plug in and get paid fairly for what they add, or they can keep selling seats to a workflow that’s going away. Either way, the direction of the integration is settled. The assistant isn’t moving into the library. The library’s going to have to start making house calls.
Sources
- Model Context Protocol — the open standard for connecting AI assistants to external tools and data sources
- CourtListener REST API documentation — public, documented API over case law, RECAP federal dockets, oral arguments, and judge data
- Free Law Project — the 501(c)(3) nonprofit operating CourtListener and RECAP
- Caselaw Access Project — archival scans of the printed reporters, served as static files
- The thirty-article verification run described above is this firm’s own work over 2026-08-18–19; its per-article source manifests and independent gate reports are maintained in the site’s repository. The vendor email quoted in reply is paraphrased from real correspondence; no vendor employee is named.
Commentary on legal technology and the business of law — the opinions are the author’s. Not legal advice, and not a recommendation for any particular firm’s research stack: what a firm can safely drop depends on its practice mix, and treatment-checking (a citator) remains a required verification step this firm currently sources separately. No client information appears in this article. Questions about anything here: Send us a message or 612-470-6529.