---
title: "ChatGPT Advertising for Law Firms: Buy the Citation - Carlos Arias"
description: "ChatGPT advertising for law firms is now permitted in the US. The ad sits below the answer. The citation sits inside it. That gap decides the budget."
url: "https://carlosarias.com/blog/ai-automation/chatgpt-ads-versus-geo-law-firm"
---

[AI & Automation](/blog/categories/ai-automation)

# ChatGPT Advertising for Law Firms: Buy the Citation

ChatGPT advertising for law firms is now permitted in the US. The ad sits below the answer. The citation sits inside it. That gap decides the budget.

  [Carlos Arias](/blog/authors/carlos-arias) · September 23, 2026  · 5 min read

![A small sponsored block sits below a paragraph of ink on warm paper, a red seal at the lower right.](/_astro/cover.BRaZftke_2akf6k.webp)

*A small sponsored block sits below a paragraph of ink on warm paper, a red seal at the lower right. AI-generated illustration by Carlos Arias .*

ChatGPT advertising for law firms went from banned to buyable this quarter, and my answer to the budget question is still no for most small firms. Not because the channel is vapor. Because the ad unit reaches a narrower slice of your buyers than an organic citation does, and the organic work has to exist before the ad is worth a dollar anyway.

Buy the citation. Rent the ad later. Here is the reasoning, with the mechanics that actually decide it.

## What changed, and how recently

For most of 2026 you could not run this even if you wanted to. OpenAI kept legal services on the disallowed list next to gambling and political advocacy, and as late as August 12, 2026 Search Engine Journal was reporting that ads for legal advice and court-related work stayed excluded while health and finance advertisers were being let in. The August revision of the ad policy moved legal into the permitted column for the United States, conditioned on the advertiser holding a valid license to practice in the jurisdiction where the ad is shown.

That is about six weeks old as of September 24, 2026.

Read the current policy yourself before anyone puts a number in a plan for you. SEJ counted four revisions to the eligibility rules in five months, the last of them dated August 10. A vendor quoting you last quarter’s rules is quoting fiction, and the vendors selling “ChatGPT Ads for Lawyers” landing pages right now are mostly doing exactly that.

## Where ChatGPT advertising for law firms sits, and where the citation sits

Nobody draws this for you, and it settles the argument.

The ad is a labeled, visually separated block at the bottom of the response. OpenAI has been explicit that sponsorship does not influence the answer text, and ChatGPT does not interleave ads into the middle of an answer the way a results page interleaves them into the listing. Search Engine Land’s breakdown of the mechanics covers the rest: a second-price auction, no conversation data handed to advertisers, and platform targeting by device. Newer formats group several advertisers into one placement instead of a single sponsored result.

The organic citation is not in that block. It sits inside the answer, as the source the model consulted to say the thing it said.

Then there is reach. Ads serve to Free and Go users only, which SEJ confirmed at the start of the test; Plus and Pro subscribers see none of it. And ads are withheld from conversations the system classifies as sensitive or emotionally vulnerable.

Read that last clause again with a personal injury practice in mind.

## The constraint most principals hit first

The conversation where somebody describes a crash at 11pm, or a firing, or a notice from immigration, is the conversation where the ad is suppressed and the citation is not.

So paid buys you research mode. Organic covers research mode and the crisis.

That is the whole constraint, and it is not a pricing objection. Pricing is actually fine. SEJ reported the ads manager offering cost-per-click bids in the $3 to $5 range, which against legal PPC is nothing. The trap is on the other end of the click. You are paying to send a pre-qualified person to a page, and if that page is the standard brochure with “Why Choose Our Firm” as the second heading, you bought an expensive bounce. Worse if the form behind it routes to an inbox nobody watches until Monday. We have written about where law firm leads actually die, and a new ad channel does not change that failure, it funds it.

## GEO for law firms is the asset; the ad is a rental

GEO, generative engine optimization, is the work of becoming the source a model quotes rather than a link it ranks. The distinction matters because the two have come apart and we can date it. Ahrefs matched 863,000 keywords against 4 million AI Overview URLs and found only 38% of cited pages ranked in the top 10 for the query that produced the citation, down from 76% seven months earlier.

Two more numbers, from Semrush’s analysis of 126 million US AI search prompts between January and April 2026. ChatGPT cites an average of 15 sources per response. And 61.7% of citations are ghost citations, where the link is present but the brand name never appears in the answer text.

Fifteen slots is a lot of room. The ad block has one, sold by auction, and you stop existing in it the day you stop paying. The citation is earned against your pages, and it keeps working in the sensitive conversations where no ad is allowed to appear. So I would not move organic budget into the ad to fund a test. Search Engine Journal’s own guidance on this is blunt about the same thing: ChatGPT Ads are not paid search, and funding them out of a channel that already produces qualified matters is a lateral move dressed as innovation.

## When I would actually test it

Not never. There is a threshold, and it is checkable in an afternoon.

- Your practice pages already get cited. You have run a prompt panel and you appear for some real question strings. If you appear for none, paid is buying reach into a firm the model cannot vouch for.
- The landing page answers a question, not a positioning statement. The shape that works is described in our piece on mining intake language for AI citations.
- Somebody answers new inquiries in minutes, not hours. The same discipline that makes missed Local Services calls so expensive applies here, and AI-referred visitors arrive further down the decision than organic ones.
- You can say what the test is for. Brand lift in one practice area, one city, a fixed 60-day spend, and a source field on the intake record that a human fills. Without that last part you will never know what the money did.
- Your bar rules clear it. OpenAI permitting an ad is not the same as your state permitting it, and the ad copy has to survive the same read as everything else you publish.

## What I will not sell you here

No ad account as the deliverable. Setting one up takes twenty minutes, and the twenty minutes is not the work. No GEO score either. Nobody has a defensible one, and I have said so in how we measure AI search visibility.

## If this is your question

The useful next step costs nothing. Write down eight questions in the words your callers actually use. Run them through ChatGPT and note two things: whether your firm appears in the answer, and whether an ad appears underneath it. Do it twice, on different days, because these systems are non-deterministic and one run tells you almost nothing.

If your firm is absent from the answers, that is the problem to fix first, before any auction. I can tell you where the ad appears and where the citation appears. Neither of us can promise the calls that follow. If you would rather have the diagnosis run by the operator who would build the fix, write us at hi@carlosarias.com and tell us which practice area you would test.

    Tags
#Chatgpt Ads
#Generative Engine Optimization [#AI Search](/tags/ai-search/)[#Law Firm Marketing](/tags/law-firm-marketing/)
#Paid Media    Share        Written by [Carlos Arias](/blog/authors/carlos-arias)

Marketing Engineer for law firms. I combine digital marketing, software, data, automation and AI to improve the whole system — from first click to signed case.

         On this page

- What changed, and how recently
- Where ChatGPT advertising for law firms sits, and where the citation sits
- The constraint most principals hit first
- GEO for law firms is the asset; the ad is a rental
- When I would actually test it
- What I will not sell you here
- If this is your question

## Continue reading

      [Law Firm Marketing](/blog/categories/law-firm-marketing) · September 21, 2026  [### Law Firm AI Search Indexing: Four Retrieval Pipeline Tests](/blog/law-firm-marketing/law-firm-ai-retrieval-pipeline/)

Law firm AI search indexing, tested in an afternoon. Four checks that tell you whether ChatGPT and Claude can retrieve your practice pages at all.

  Carlos Arias · 7 min
      [Law Firm Marketing](/blog/categories/law-firm-marketing) · September 20, 2026  [### Measuring AI Search Visibility for Law Firms: Your Report Still Leads With Rankings](/blog/law-firm-marketing/ai-search-measurement-law-firm-marketing/)

Measuring AI search visibility for law firms starts where the ranking report stops. What to track, what to drop, and the one number nobody can attribute.

  Carlos Arias · 7 min
      [Law Firm Marketing](/blog/categories/law-firm-marketing) · September 19, 2026  [### Your Law Firm Can Get Sued Over Its Own Website (The Widget Won't Save You)](/blog/law-firm-marketing/your-law-firm-can-get-sued-over-its-own-website-and-your-accessibility-plugin-won-t-save-you/)

Your law firm can get sued over its own website. The ADA names lawyers' offices by statute, the overlay widget is not a defense, and the fix pays twice.

  Carlos Arias · 8 min

## Stay in the loop.

One email when it’s worth it — new posts and updates, no spam.

Thanks — check your inbox to confirm.

Free. Unsubscribe in one click.
