Your Software Just Became an AI Company. Here's What to Ask Your Vendor.

On Oct. 26, 2026, Clio, the practice management company many solo and small firms use, unveiled a new logo and brand and now calls itself a legal AI company. The logo changes nothing by itself, but when your vendor calls itself an AI company, expect more of the software you already pay for to involve AI, and AI features touch client data. Send your main vendor five questions in writing: is our data used to train AI, where is it stored, how do we delete it, which AI features are on by default and can an administrator turn them off, and which model providers sit underneath the product and do they keep anything we send. Keep the answers in your file.

Where does this come from?#

This guide is the written companion to Episode 4 of our podcast, "Your Software Just Became an AI Company". The episode covers legal AI news from Oct. 26 to 27, 2026. Every fact here comes from the sources listed in the episode's show notes. The episode also covers judges in one state trying an AI research tool, and a proposed court rule in another state on serving some court papers by email, with public comments due before Jan. 15, 2027. This guide sticks to the vendor question, because it applies to every firm that pays for software. The advice sections are our practical suggestions for solo and small firms, not rules and not legal advice.

If you'd rather listen, the episode page has the audio and the full transcript.

What did Clio announce?#

On Oct. 26, at ClioCon 2026, its annual conference, Clio unveiled a new logo and a new brand. In its announcement, it now calls itself a legal AI company, not just a practice management company.

Practice management software is the system a firm uses to run matters, calendars, documents and billing. For a lot of small firms, it's the filing cabinet and the front desk in one. That's why a change in how this particular kind of vendor describes itself deserves a second look.

Clio says it serves more than 400,000 legal professionals in over 130 countries. Those are the company's own figures. Clio also says its commitment to solo and small firms remains.

We covered the announcement in our news item on the rebrand, and our guide to what Clio's rebrand changes for your firm covers what the announcement does and doesn't say.

Does a new logo change anything at your firm?#

Not on its own. You log in on Monday and it's the same software.

What the logo does is tell you where a company is putting its effort. When your main vendor calls itself an AI company, expect more of the product to involve AI over time. And AI features touch client data, meaning the documents and notes your firm keeps in that system.

So treat the rebrand as a reminder, not an emergency. It's a good excuse to reread the fine print on the system that holds most of your client information. And the questions below aren't about Clio in particular. They work for any vendor whose software now includes AI, whether that vendor announced it with a new logo or added it in a routine product update.

What this means for you: Before more AI features arrive in the software you already use, get written answers from your vendor about what happens to your client data, and keep them where you can find them.

What should you ask your vendor?#

The episode starts with three questions from our guide to AI tools for small law firms, then adds two more. Here they are, with what each one is really asking.

1. Is our data used to train AI? Training means the vendor uses your documents to teach its AI. Ask the question broadly. You want to know about the files you upload and also about what your staff type into the AI features.
2. Where is our data stored? This is a plain location question, so ask for a plain answer. Ask too whether it covers every company that handles your data on the vendor's behalf, not just the vendor itself.
3. How do we delete it? Ask how deletion works while you're a customer, and what happens to your data if you leave. Better to learn the answer now than on the way out.
4. Which AI features are on by default, and can an administrator turn them off? The episode's point here is that this varies by vendor. An administrator is the person at your firm who controls account-wide settings. If nobody at your firm knows who that is, find out before you send the email.
5. Which model providers sit underneath the product, and do they keep anything we send? More on this one below, because it's the easiest one to overlook.

Why ask who built the AI?#

Because you can't assume the AI inside your software was built by the company you pay.

Two definitions help here. An AI model is the engine that reads and writes text inside an AI tool. A model provider is the company that builds the underlying AI model a software product uses. Your vendor builds the product you actually log in to. The model provider may be a different company: the company behind the company.

The episode used a piece of business news to make the point. On Oct. 26, Anthropic, an AI model developer, said it filed a registration statement with the Securities and Exchange Commission for a proposed initial public offering. Anthropic says it plans to list on the Nasdaq under the ticker symbol "ANT." As of Nov. 5, 2026, Anthropic's announcement says the number of shares and the price range haven't been set.

In plain English: the Securities and Exchange Commission (SEC) is the U.S. agency that oversees the sale of stock. A registration statement is the paperwork a company files with the SEC before selling stock to the public. An initial public offering (IPO) is the first time a company offers its shares to the public. Nasdaq is a U.S. stock exchange, and a ticker symbol is the short code a stock trades under.

That announcement is about stock, not products, and it isn't a reason to do anything with your software. It's a reminder that the company that builds an AI model and the legal software brand you see on your screen don't have to be the same company.

Here's why the fifth question matters in practice. If your vendor tells you it doesn't keep your data, that answer may only describe the vendor. Ask whether the same is true for every model provider that touches what you send. You want one answer that covers the whole chain, not just the first link.

How do you send the questions?#

Start with whichever system holds the most client information. For many small firms, that's the practice management software. If you have time for only one email this week, send it there.

Send the questions in writing. Use the support address or your account contact, and ask for a written reply. A phone call is fine for context, but the answer you'll rely on later is the one you can print.

Here are our suggested questions, worded so you can adapt them. They're a starting point for a support email, not legal advice and not contract language:

"We're reviewing how the AI features in your product handle client data. Please answer in writing: (1) Is our data, including documents and anything our staff type into AI features, used to train AI? (2) Where is our data stored? (3) How do we delete it, during the contract and after we leave? (4) Which AI features are on by default, and can our administrator turn them off? (5) Which model providers does your product use, and do they keep anything we send? Where your agreement or data terms already answer a question, please point us to the section."

We suggest that last sentence because a sales answer and a contract term are not the same thing. Asking the vendor to point to the section tells you which one you're getting.

When the answers come back, save them in the file you keep for that vendor, with the date. If a client ever asks how you protect their information, you can show them the vendor's own words instead of describing them from memory.

How should you read the answers?#

Slowly. Here are the distinctions we suggest keeping in mind.

Not training on your data and not keeping your data are different promises. A vendor can truthfully say it doesn't use your documents to teach its AI while still storing what you send for some period. If the answer only addresses one, ask about the other.

"Can be turned off" isn't the same as "is off." If a feature is on by default, someone at your firm has to actually change the setting. After you get the answer, have your administrator check the settings and note what they changed.

A vague answer is a reason to ask again. If the reply says something general about taking privacy seriously, write back with the specific question it didn't answer. You're the customer, and these are reasonable questions.

Answers can go stale. A vendor that's adding AI features will keep adding them. When your vendor announces something new, it's worth asking the questions again.

What about the ethics rules?#

A quick definition first. Generative AI is software that writes something new, like a draft or a summary, when you type a question or an instruction, rather than just finding a match. It can sound confident and still be wrong. When it confidently makes something up, people say it hallucinates.

When you use it, four common ethics duties come up. Check how your state's rules handle each one. Here's how each connects to the questions above:

• Competence: know roughly what the tool does, and check its work. Knowing which features are on, and whose model runs them, helps with that.
• Confidentiality: keep client information safe. The training, storage, deletion and model provider questions help with that.
• Client communication: tell clients what they need to know. Written vendor answers make that conversation easier, and more accurate.
• Reasonable fees: keep what you charge reasonable, whatever tools you use to do the work.

Check your state's ethics rules and any bar guidance on AI before you roll this out. This isn't legal advice. Your state's rules control.

What if your firm runs on several systems?#

Then the same questions go to each vendor, in order of how much client information each one holds. The episode was clear that this is a list for every vendor, not just the one that rebranded.

As one of the hosts put it: "At some firms, the software list is email and good intentions." If that's your firm, the list is short. Your email provider still counts, though, and so does any tool that staff paste client information into.

Frequently asked questions#

Do these questions only apply to practice management software?#

No. They apply to any vendor whose software includes AI features. Start with the system that holds the most client information and work down the list.

Does an AI model developer going public affect my firm?#

Not directly. The announcement is about stock, not products. It's a reminder that the AI inside your software may come from a different company than the one you pay, which is why you ask which model providers your vendor uses and whether they keep anything you send.

Is this legal advice?#

No. This guide summarizes the sources behind Episode 4 of our podcast and offers general practice suggestions. Ethics rules and bar guidance on AI vary by state, so check yours.

Next step#

Listen to Episode 4, "Your Software Just Became an AI Company" and read the full transcript. Then send your main software vendor the five questions and file the answers. For more episodes, see the podcast page, or get the written version in our newsletter.

Check your state's ethics rules and any bar guidance on AI before you roll this out. This isn't legal advice. Your state's rules control.

This guide is general information, not legal advice. Facts come from the sources listed in the show notes for Episode 4 of our podcast; figures from Clio are the company's own. Some listings on AI Tools For Law may be sponsored or include affiliate or referral links, and paid placements are always labeled "Sponsored"; see our disclosure. None of the links in this guide are paid placements, and nothing in the episode is sponsored.