# How often should a law firm check its AI visibility? _By Apoorv Khanna, September 29, 2026_ Start with a weekly review of your firm's AI visibility. Ask a consistent set of client questions, compare each service separately and save the answers. Weekly is a practical operating rhythm, not a scientifically established interval or a promise that every change needs a response. If you automate collection, daily observations give you more material to review on Monday. Keep the decision-making cadence slower than the collection cadence. One unexpected answer is usually a reason to inspect the test before rewriting a practice page. ## Keep the test stable enough to compare Choose questions that fit the work your firm accepts and the market it serves. Include the city in the question. Use client language about a legal need, such as finding a car-accident lawyer, rather than asking whether your firm's name sounds reputable. Save the exact wording. Record the service, date, available model or search mode, location settings and whether an answer returned. Use a fresh conversation for each comparable test so you do not deliberately feed the answer with a prior discussion of your firm. You cannot hold every part of an external AI service constant. You can record the conditions you control. If you change a prompt or add a new practice area, start a separate series for it. Our [guide to checking local recommendations](/blog/which-law-firms-chatgpt-recommends) walks through the initial setup. ## Review mentions, accuracy and sources together A mention count is a starting point. Read the answer that contains the name. Check whether it describes the correct office, practice area and language support, and whether the linked website belongs to your firm. An appearance with incorrect facts needs different work from an absence. Then inspect the cited pages. A newly cited directory, a revised firm profile or a different article gives you something concrete to review. Record what changed in the answer and what changed on your own pages as separate observations. Keep the denominator visible. “Named in four answers” means little without knowing how many questions returned usable results. A failed check, an answer with no firm recommendations and an answer that recommends competitors are three different outcomes. Do not collapse them into the same zero. ## Keep a separate view for each AI In our [September 27 Austin sample](/blog/ai-personal-injury-lawyer-austin), ten questions sent to three services produced 27 returned answers and named 57 firms. Seven firms appeared across all three services somewhere in that run. FVF Law Firm appeared in ChatGPT and Perplexity answers but not in the captured Google AI answers. Those results show why an aggregate score can hide a useful difference. They do not show how any firm's visibility changed over time: the sample covers one day. To see a pattern over time, you need repeated observations under comparable conditions. Put the question and service next to every result. A drop in one combination can then be investigated without treating it as a disappearance everywhere. ## Log edits without assuming they caused the next answer When you correct a directory profile or revise a service page, record the date and the change. At the next review, inspect whether the edited page appears among the cited sources and whether the answer describes your firm more accurately. An improved answer after an edit is useful to observe, but timing alone does not prove causation. Other pages, competitors and the service itself can change. Repeated results make a pattern easier to assess; they still do not reveal a private ranking formula. Assign the review to one person and give them a short task list: check incomplete runs, read changed answers, verify factual errors and choose the next correction. Keep the saved responses so a colleague can inspect the same evidence. Avoid turning each weekly report into a requirement to publish another generic article. ## Let daily collection feed a Monday review [AEO Monitor's law-firm recipe](https://vaaya.ai/aeo-monitor/law-firms) offers a free beta check of up to five questions across ChatGPT, Perplexity and Google's AI Overviews. You can enter a website or describe your practice and city without creating an account. The recipe describes an optional email report built around daily checks and one Monday summary: how often the firm was named, which competitors appeared, cited sources and changes. Review the underlying answers, choose a correction you can substantiate, and put the next review on the calendar. ## Questions **How often should a law firm review AI recommendations?** A weekly review is a practical starting point. Keep the same questions and conditions so comparisons mean something. Daily automated collection can provide more observations, but it does not require daily changes to your website. **Does one missing AI mention mean visibility has fallen?** Not necessarily. Check whether an answer returned, whether the question and settings changed, and whether the absence repeats. A failed request or missing AI Overview should be recorded separately from an answer that names other firms. **What does AEO Monitor's Monday report cover?** The recipe offers daily checks with a weekly summary of named firms, competing mentions, cited sources and changes. Email is optional for receiving the report; the initial beta check does not require an account.