By matching the directories, review sites and your own site on the same name, address and practice areas, then giving AI a page that answers the buyer's question directly. Local legal queries get answered from whichever sources agree with each other, not from the firm with the nicest website.
This is the local, professional-services version of a question every business owner now asks: how do I get recommended by ChatGPT. The mechanics are the same five steps, with one addition that matters more for a law firm than for most businesses: the directories and review platforms a buyer already trusts for "which lawyer should I hire."
By pulling together whatever it can find on legal directories, review sites, Google's local data and the firm's own site, then naming the firm those sources agree on most. Ask an engine with web search on, "best employment lawyer in Stockholm," and watch which sources it opens: usually a mix of a directory listing, a review aggregator and one or two firm websites. A firm that is missing from, or inconsistent across, those sources gives the engine less to agree on, and less reason to name it over a competitor who is consistent everywhere.
Why it matters: you are not competing only on your own website. You are competing on every page that mentions your firm, most of which you do not control directly.
Because a mismatch is the easiest signal an engine has that two listings might not be the same business. If your website says "Nordström & Lindqvist Law" and a directory lists you as "Nordstrom Lindqvist Advokatbyrå" at a different suite number, an engine reconciling sources has to guess whether it found one firm or two. The fix is mechanical, not creative: pick one name, one address format and one phone number, and use that exact string everywhere your firm is listed.
Do this: run the free external profiles check (check 4.2) to see where your listings currently disagree.
Yes, and a complete one. For a "near me" or city-specific legal query, Google's own local results and AI Overviews draw directly on Business Profile data, and the same name, address, hours and practice areas you set there are what review sites and directories tend to echo back. A profile left unclaimed, or claimed and abandoned with an old address, is one more place your firm's identity can drift out of sync with itself.
Rule of thumb: treat your Business Profile as a source of truth other listings should match, not a one-time signup task.
An Organization block with your legal name and sameAs links, and LegalService schema on your practice pages, both matching the visible text exactly. Schema.org defines LegalService as a type of local business built for exactly this: a named provider, an address, and the service it offers. Structured data does not make an engine trust you more on its own; it only removes ambiguity if the markup says the same thing a visitor reads on the page.
Proof: schema.org — LegalService · Google — Understand how structured data works.
Check it: the free schema and entity check flags where your markup and visible copy disagree.
Yes. A page titled "Employment Law" that opens with a firm history gives an engine nothing to quote for "what does an employment lawyer do in a wrongful termination case." Open each practice page with the plain answer to the question a buyer in that situation is actually asking, in the first sentence, then add the firm's experience and process underneath. An engine summarizing a buyer's question can lift a direct answer; it has to work much harder to extract one from a page that buries it.
| What AI checks | Fast to fix (days) | Slow to fix (weeks to months) |
|---|---|---|
| Identity match | Correcting your name/address/phone on your own site | Getting a directory or review site to correct its copy of it |
| Structured data | Adding Organization and LegalService schema | — |
| Page content | Rewriting a practice page to answer first | — |
| Third-party trust | — | New reviews, new directory listings, press mentions |
| Entity records | — | A sourced Wikidata item, if a reliable source exists |
Worth trying, but only if a source Wikidata considers reliable already mentions your firm, such as a bar association listing or local press. A Wikipedia article needs significant independent press coverage that most firms will never have; a Wikidata item has a lower bar and gives engines one more clean, sourced answer to "who is this firm." The full case for it, and what counts as a reliable source, is in why Wikipedia and Wikidata quietly decide what AI knows about you.
Everything that depends on someone else: new reviews, a new directory listing, press coverage, a Wikidata item. Those build over months, not days, because they need another site or another person to act. Identity matches, schema and page rewrites are entirely in your own hands and can go live this week. Do the fast fixes first; they are also the ones a mismatched listing elsewhere can undo if you skip them.
None of this requires buying anything, and most of it takes under an hour:
We can't confirm a direct read, but a complete, accurate profile shapes Google's own local results and AI Overviews, and the same name, address, phone and hours tend to show up again on the directories and review sites that ChatGPT, Claude and Perplexity do cite.
Yes, more easily for a specific question than a generic one. A firm that is one of five results for "employment lawyer in Stockholm" on the directories and review sites AI reads has a real shot; one competing for "best law firm" nationally does not.
Matching your name, address and phone number everywhere it appears, completing your Google Business Profile, and adding LegalService and Organization schema to your site. All three take an afternoon and none of them wait on anyone else.
Almost certainly not yet. Wikipedia requires significant independent press coverage most firms do not have. A Wikidata item is the realistic target, since it accepts an entity backed by a reliable source, not sweeping press coverage.
Access and schema fixes can reach an engine with live search within days of the next crawl. Directory listings, reviews and a Wikidata item build over weeks to months, since they depend on other sites and, in some cases, other people.
The free check reads your site the way ChatGPT, Claude and Perplexity read it, and shows who gets named instead. 3 minutes, evidence included.