0
firms in common when we asked two assistants the same Dallas car accident question. Six names, no overlap.
Picked by Agents runs, Meta Muse and Instinct, Sept 2026
48%
of local queries got completely different answers from four major AI engines. They fully agreed on only 4%.
Yext, Feb to Aug 2026, 491,972 queries
42%
of AI citations for lawyers point to Google profiles. Avvo gets 2.5% and Justia 3.2%.
BrightLocal AI directory sources, Aug 2026
The crash was yesterday. The question goes to an assistant.
We asked two assistants the question an injured person actually types. I got rear-ended on I-35 in Dallas yesterday and my neck really hurts, who should I call, find me the three best car accident lawyers. Each one searched, read firm sites, checked the State Bar and review directories, and came back with three firms and a reason for each.
The two lists had no firm in common. That is not noise. It is what an open market looks like before anyone has worked out what the assistants read. Your billboard and your search ads reach the people who were going to search anyway. The shortlist reaches the ones who asked a machine to decide who they call first.
“Six firms named, none named twice. The shortlist in personal injury is still up for grabs.”
What the assistants looked for
Reading every reason both assistants gave, the same kinds of evidence decided who made the list and who was rejected.
- 01
A page for the incident
The strongest pick in one run had a page for rear-end collisions in Dallas naming the lawyer who handles them. Firms that only had a general car accident page, or none, were described in vaguer terms.
- 02
Standing the assistant can check
State Bar profiles showing the lawyer is eligible to practice with no public discipline, and board certification in personal injury trial law, were cited by name.
- 03
Independent reviews
Avvo ratings, FindLaw reviews and review volume on aggregators were quoted directly. A firm with only a handful of reviews was rejected as too thin a track record.
- 04
Fit for this kind of case
Firms focused on criminal defense, medical malpractice or catastrophic cases were rejected for a routine rear-end collision, however decorated. So was a firm whose listed phone number put it in another city.
- 05
What neither assistant could find
Fee percentages were unpublished at every firm, and neither assistant could confirm when a consultation was available or whether the firm would take the case. Every path ended at a call or a form.
What wins a place on the shortlist
The firms that got named made the assistant's job easy. A page for each type of incident in each city they serve, naming who handles it. Bar standing and certifications stated plainly and linkable. Reviews on the directories the assistants actually read, consistent with the Google profile. Clear terms for a free consultation, and ideally a way to see when one is available.
None of that requires a bigger ad budget. It requires publishing, in plain text on your own site, the facts an assistant uses to decide whether you fit the case in front of it.
What we do for a personal injury firm
- 01
Run the injured person's real questions, monthly
Car, truck, motorcycle and premises questions in each city you serve, through the assistants people use, against the firms that beat you. You see who got named and, in the assistant's own words, why.
- 02
Incident and city pages that answer the question
One page per case type and office, naming the lawyer, the court experience and the consultation terms, written for an assistant to cite.
- 03
Standing and reviews the assistant can confirm
Bar profiles, certifications, Avvo and the other directories the assistants read, made consistent with your site and your Google profile.
- 04
A consultation an assistant can see
Where you are willing, published consultation times or a request path that does not end at a callback form. On Picked we tag AI-sourced calls and forms in your intake so you can see which cases the assistant sent.
- 05
Within the rules, every time
Everything we publish respects your state bar's advertising rules. No "best" claims, no promised outcomes, no results presented without the context your bar requires. We give you the copy and you approve it before it goes live.
Who this is for
Plaintiff personal injury firms with one office or twenty, where a founding partner or a marketing director owns the intake numbers. You already spend heavily to be found. Even the vendors selling to you pay $143 a click on "personal injury lawyer advertising", per the Google Ads keyword data we pulled in September 2026.
Legal is not an agentic category yet. No assistant books a consultation or signs a client. We sell a place on the shortlist and proof you are on it, not cases.
Personal injury advertising as usual vs Picked by Agents
| Personal injury advertising as usual | Picked by Agents | |
|---|---|---|
| Where the client decides | A billboard, a TV spot, a page of search ads | One answer with about three firms and a reason for each |
| What it measures | Impressions, clicks, cost per signed case | Whether the assistant named your firm, and why, every month |
| What it fixes | Budget and creative | Incident pages, bar standing, reviews, consultation terms |
| Compliance | Reviewed by the firm after the fact | Written to your state bar's advertising rules, approved by you first |
| Proof | A media report | The assistant's own words, then AI-sourced calls in your intake |
| Price | Media spend that stops working when it stops | $2,997 or $4,997 a month, month to month. The pages stay yours. |
Questions we get asked
- How many injured people actually ask an assistant for a lawyer?
- Nobody has measured it for legal specifically, and we will not invent a number. What is measured is that 45% of US consumers have used an AI tool for local business recommendations, per BrightLocal's 2026 survey, and that professional services are where people take longest to decide. We start with your own intake data and add AI-sourced calls as we tag them.
- Why did two assistants name completely different firms?
- They searched differently and weighted different evidence. One leaned on State Bar records and incident-specific pages, the other on review volume and directory ratings. Yext's study of nearly 500,000 local queries found four major engines fully agree only 4% of the time. That is why we run a panel of assistants, not one.
- Can you make us the best rated firm in the assistant's answer?
- We do not write "best" claims or promise results, and neither should you under most state bar rules. We publish the facts the assistants weigh, accurately, and let them draw the conclusion. What we guarantee is that you are named for the questions we agree, not how the assistant ranks you or what any case is worth.
- Do we have to publish our contingency fee?
- No. It is your call, and it has to be accurate and comply with your bar's rules if you do. Neither assistant found a fee percentage at any Dallas firm, so a firm that states its standard terms clearly would stand out.
- Does this replace our ad agency?
- No. Keep your media and your agency. We fix the one thing they do not measure, whether the assistant names you when someone asks who to call, and the incident pages we build usually help your paid landing pages too.
See which firms the assistants name in your city.
The report is free. It shows who they named for the injury questions your clients ask, and what they found or could not find about you. Then we fix it, guaranteed.
Or write to Greg at greg@gregmiller.ai.