The questions people ask AI assistants to find a lawyer, by metro
These are questions, not answers. Pick a metro and you get the questions a real client would type into ChatGPT, Gemini, or Perplexity to find a lawyer there — with the county names, the metro nickname locals actually use, and the state-law phrasings a firm would never think to write down. Copy a pack, or download it as CSV or plain text. No signup, no email.
Nothing in a pack is a measurement. No pack reports what any assistant said, about any firm, anywhere. The point is the input side: if you are going to check whether assistants recommend you, these are the questions worth checking with.
25 markets, 18 states
One page per metro. The practice area named beside each is the pack that opens first; the others are on the page, collapsed.
California
Florida
Texas
What a pack is
- A list of client-voice questions built by slot-filling one metro's verified catalog vocabulary into question templates. No model call, no scraping, no I/O.
- Deterministic: the same catalog and templates produce the same list on every build, so two people running a pack a month apart run the same pack.
- Fully labelled. Every question carries its intent family, its phrasing mode and the exact location string used, so you can slice the list instead of taking it whole — the CSV download carries those columns with it. It also carries a priority — our own weighting of the intent families, which decides how many questions each one gets. That one is an editorial choice, not a measurement of what anyone asks.
What a pack is not
- Not a measurement. Nothing here says what an assistant answered, or which firms it named.
- Not a keyword list, and not a promise. Asking these questions — or publishing pages that answer them — does not make an assistant recommend you. We have not measured that, so we do not claim it.
- Not a claim about the market. Which practice areas get a pack is our editorial choice about what to render, not a finding about what local firms practise.
Why the packs read differently from each other
Because the vocabulary is real. A Houston pack says Harris County and asks how comparative negligence works under Texas law. A Tampa pack says Hillsborough County and “the Tampa Bay area”, and asks whether a no-fault state still needs a lawyer. A Charlotte pack carries contributory-negligence phrasings that cannot appear in most other markets, because North Carolina is one of the few states that still works that way. None of that is decoration bolted onto a template — it is the substance of the pack, and it is why one metro's list is not another metro's list with the city name swapped.
Each pack also shows its own reach. Every market page lists the full catalog record — city, counties, nickname, suburbs — and marks which of those strings the pack actually used, because a pack of this size works down from city and county language and does not get to the end of the list. You should be able to see the edge of the thing you are being handed.
Where a state has no quirk we can verify, the packs for its metros simply have no state-law section. There is no “standard rules apply” filler anywhere: an omitted field in our catalog means unknown, and unknown gets no sentence.
Before you use one
Two of the ten intent families produce questions in the voice of someone who has just been hurt, arrested, or bereaved. They are in the packs because that is how the question actually reaches an assistant, and an honest measurement has to include them. They are measurement input, not a list to write landing pages against.
And the standing caveat, which applies to anything you find by running these: being named by an AI assistant says something about coverage and about which sources the assistant leaned on. It says nothing about the quality of an attorney's representation. If you want the rules we hold ourselves to when we run packs like these at scale, they are on the methodology page; the metric that comes out the other end is defined on Case Recommendation Share; and the longer arguments — including why identical reruns disagree — are in the writing.
Run three of these questions across five assistants — free
Working a pack by hand takes an afternoon. The free Briefly check does three questions across five assistants and shows the actual answers: whether your firm is named, which firms are named instead, and which sources shaped the answer.
Check your firm free →No signup. No credit card. No sales call — and the result is not held behind an email address.
Briefly measures what AI assistants say. It does not rank, rate, or endorse attorneys, and nothing here is legal advice. All free tools