Why family law is where this bites first
High-consideration legal categories are among the most AI-saturated searches there are, and legal buyers increasingly start in a chatbot rather than a search bar. The buying-intent question — "who should I hire" — is now answered with a short list of named firms, assembled from sources most attorneys have never audited. Our own family-law panels have found the same pattern in market after market: the established independent firm is recommended only when a prospect searches its own name, and is absent everywhere else. That contrast — reputable on demand, invisible in the recommendation — is the exact gap we close.
What actually moves an engine (from our panels)
Every engine has a different lever, and generic "SEO" moves none of them reliably. What we've watched work, engine by engine:
| Surface | What decides the recommendation | The lever |
|---|---|---|
| ChatGPT | Directory depth — Avvo and Expertise profiles carry the shortlist | Complete, well-reviewed directory profiles |
| Reasoning tiers | Verifiable credentials — bar standing and board certification | Bar-profile hygiene; specialty certification surfaced |
| Claude | Google review score, volume, and what reviews say about fees | Earning reviews and fee-friendly review themes |
| Perplexity | Vertical directories and whether your pages are quotable | Answer-first pages with extractable facts |
| Google AI | GBP, review sites, and your own educational library | Concrete-fact pages: pricing, process, timelines |
The single strongest pattern we've measured, across engines: the firm that publishes real numbers wins the cost questions. When a prospect asks what a custody lawyer costs, the engine quotes the one firm that put its fees on a page — and names that firm because of it. Most firms publish nothing. That's an open answer, waiting.
Why the approval gate is a compliance feature, not a nicety
Every state bar holds the lawyer responsible for every marketing claim — including ones generated by AI, and including statements that mislead by omission. California is the sharp edge of where this is heading: SB 37, in force since October 2025, defines an advertisement as any written, recorded or electronic communication, gives you 72 hours to pull a non-compliant ad from electronic media, and exposes you to statutory damages from $5,000 to $100,000 per unique advertisement. SB 574, still moving through the legislature, would bar delegating the practice of law to generative AI and require you to verify every AI-produced citation you file. Responsibility for a vendor's words is not an SB 37 rule — it comes from your own conduct rules on supervising non-lawyer help, and it lands the same way. Your state's rules may be lighter today, but the direction is the same — and a robot that publishes to your site with no drafts to approve is hard to reconcile with that duty.
Nameworthy is built the opposite way. Nothing reaches your site until you approve it, and that approval is documented:
- A per-page approval log — who approved, when, which version.
- A versioned archive that supports record-retention requirements.
- An instant-unpublish path for anything you need down fast.
- A responsible-attorney block on every page, and banned-phrase checks in drafting that flag guarantee language, uncertified "specialist" claims, and testimonial wording before you ever see it.
The reputation risk you can't see
When a prospect asks whether your firm is reputable, you don't get to choose which version of your firm the engine serves. We've documented answers that quietly surfaced a real discipline record on one run and omitted it on the next, and engines that embellished credentials no one can verify. Branded-query monitoring is a real, demonstrable deliverable, and an intake reputation check belongs at the start of any engagement.
How we measure
Six surfaces, every buying-intent question run repeatedly, every run on file. Read the method →
See your firm's numbers
A free report on your real client questions, in 48 hours. Get your report →