Half of Local Searches Now Run Through AI — And It Doesn't Read Your Website First
The number that should stop every managing partner this month comes from BrightLocal: the share of people who use AI to help find or choose a local business jumped from 6% a year ago to 45% today. In twelve months, asking an assistant to recommend a nearby firm went from a fringe behavior to something nearly half the market does. For an immigration practice, that means a family looking for an H-1B or asylum attorney is increasingly typing the question into ChatGPT, Copilot, or Google's AI Mode rather than scrolling a page of blue links. And here is the part most firms miss: when AI answers that question, it does not start with your website. It assembles a recommendation from the surfaces around your website — your map profile, your directory and review listings, and Bing Places — and only then, sometimes, reaches your site to confirm.
That reframes the whole problem. A firm can have a beautiful, complete homepage and still be invisible at the moment of recommendation, because the assistant made its shortlist from records the firm has never opened. Visibility is no longer one page you control. It is a set of surfaces you populate but don't own, and they only work in your favor when they agree.
The Map Profile Is the First Record AI Reads
For local-intent queries — "immigration attorney near me," "green card lawyer in [city]" — the map profile is often the first structured record an assistant pulls. It is a clean, machine-readable summary of what your firm does, where, and for whom, and AI leans on it precisely because it is structured. The catch is that AI can only recommend from the fields you've actually filled in.
A primary category set to "Lawyer" instead of "Immigration attorney" tells the model you're a general practice. A blank services section gives it nothing to match against a client asking about removal defense. A description that never says H-1B, asylum, or naturalization is a description the model can't connect to the words clients type. The profile is doing recommendation work every day, whether or not anyone at the firm has looked at it since the year it was created.
Bing Places Is the Surface Nobody Claimed
Copilot — Microsoft's assistant, embedded across Windows, Edge, and Office — draws its local knowledge from Bing Places, not Google. And Bing Places is the listing almost no immigration firm has ever claimed. Microsoft often auto-generates a profile from whatever it scraped years ago, and that unverified record then answers on your behalf.
The result is a surface that can quietly contradict everything else: an old suite number, a category that no longer fits, a phone line that was disconnected two moves ago. When a prospect asks Copilot for an immigration attorney and the assistant reads a stale Bing profile, it either passes your firm over or recommends it with wrong information. Claiming and correcting that listing is often the single highest-return hour a firm can spend, precisely because the bar is so low — most competitors haven't done it either.
Directories and Reviews Are the Shortlist AI Trusts
Beyond the map and Bing, assistants concentrate their trust in a short list of third-party platforms — legal directories like Avvo, Justia, FindLaw, and Martindale, plus review profiles and state-bar listings. These are the sources AI treats as independent confirmation that your firm is real, active, and practices what your site claims. You cannot add yourself to that trusted shortlist by publishing more blog posts. What you can do is make sure that where those platforms already describe your firm, the description is accurate, current, and consistent with everything else.
A firm listed correctly across the directories — same name, same practice areas, same address, with recent reviews — is one an assistant can recommend with confidence. A firm with a half-built Avvo page, a directory entry naming a partner who left in 2022, and no review newer than last year is one the model will quietly skip in favor of the practice next door whose records line up.
Consistency Is the Multiplier
The reason all of this matters more than any single listing is that the surfaces reinforce one another when they agree and confuse the model when they conflict. AI is trying to decide whether three records describing "Garcia Immigration Law," "Garcia & Associates," and "Law Office of Maria Garcia" are one firm or three. A mismatched suite number, a maiden name on one profile, a different phone on another — each is a seam where the model loses confidence and splits your reputation across records it never merges. The firm that wins the recommendation is usually not the one with the fanciest site. It's the one whose name, address, phone, and practice areas read identically everywhere the assistant looks.
What to Check This Week
The point of multiple surfaces is that no single fix covers them, so here is the order to work through them:
- Open your map profile and read it as an AI would. Confirm the primary category is "Immigration attorney," the services name real work, the description uses the words clients type, and the hours and phone are current. Fix the weakest field now.
- Claim your Bing Places listing. If you've never done it, this is likely your biggest blind spot. Verify the category, services, and NAP match your map profile exactly.
- Audit your directory and review listings. Check Avvo, Justia, FindLaw, Martindale, and your state-bar directory for accuracy, consistent practice areas, and a firm name that matches everywhere. Retire any entry naming a former partner or old address.
- Reconcile every surface against the others. Make your firm name, address, phone, practice areas, and service cities identical across the map profile, Bing, directories, and your site — so the model reads one firm, not three.
The Bigger Shift
For a decade, being found online collapsed into a single question: where does the firm rank? A jump from 6% to 45% in a single year makes clear the question has changed. Nearly half of prospective clients now let an assistant assemble a recommendation for them, and it builds that recommendation from a map profile, a Bing listing, a handful of trusted directories, and only then your website. A managing partner who watches only the website is watching the last surface AI reads, not the first.
None of this requires an agency or a new budget line. It requires knowing which surfaces AI reads before it recommends anyone, and making each one accurate and consistent. The firms that do will be the ones AI names to the next family searching for help.
Clientory helps immigration law firms become visible to the families and individuals who need them most — turning AI search into consultations, and consultations into clients. See how your firm shows up across every AI surface — get your visibility score at APP.CLIENTORY.ORG.