ChatGPT Ads · OpenAI Ads Manager · Paid AI Placement
ChatGPT Ads for Lawyers & Law Firms
Legal services became eligible for ChatGPT advertising in August 2026. This page covers what OpenAI's policy actually permits, the placement rule that keeps ads out of a large share of claimant conversations, what the channel is worth testing at, and where it sits next to earning the citation instead of buying the slot beneath it.
Quick answer
Yes, US law firms can run ChatGPT ads. OpenAI's ad policy permits ads for legal advice, representation, and legal services in the United States when the advertiser is licensed to practice law in the jurisdiction where the ad is shown, and it names personal injury, legal claims, immigration, and document preparation as included. Legal was added in policy version 1.5, August 2026, reversing an earlier ban, which is why much of the published guidance on this topic is still wrong. Three constraints matter more than the permission: approval in the legal category is manual and case-by-case, ads reach only Free and Go tier users, and ads are blocked from sensitive conversations — including personal health and sensitive user journeys, which is where a great deal of plaintiff intake conversation actually happens. For most mass tort and PI firms the channel is a funded test, not a replacement for earning the citation organically.
Policy positions on this page reflect OpenAI's Ad policies as updated September 10, 2026 (v1.6). This category has changed six times in seven months. Verify before you budget.
How the policy got here: six revisions in seven months
The contradictory advice you will find on this topic is mostly a dating problem. Pages written before August 2026 say law firms are banned, and they were right when they were published.
Initial publication of the OpenAI Ads Policies. Legal services sit outside the launch categories, which are limited to consumer verticals.
Placement policy refined. Medical, legal, and financial advice contexts are no longer categorically blocked from carrying ads. This changed where an ad may appear, not who may buy one.
Ad integrity section added, covering how ads are reviewed, monitored, and enforced. Around the same time, and separately from the policy, OpenAI opened its self-serve Ads Manager in beta.
Advertiser policies introduced. Financial and health services sections clarified to list eligible categories and markets. Legal still not permitted.
Housing and job listing policy clarified.
Legal services permitted in the US. This is the change that opened the channel to law firms.
OpenAI clarifies its right to decline ads that conflict with its advertising principles, business interests, or competitive position.
One wrinkle worth knowing, because it is the source of most of the confusion: OpenAI's policy page still carries a legacy summary sentence listing legal services among categories disallowed at launch, several paragraphs above the detailed Legal services section that permits them. The detailed section and the changelog are the operative text. Anyone quoting the summary line at you is quoting a stale sentence in a live document. For how the policy plays out case type by case type, see ChatGPT ads for personal injury lawyers.
What qualifies your firm, and what still gates you
Eligibility under the written policy is necessary and not sufficient. These are the six things that decide whether a campaign runs.
Licensed in the jurisdiction where the ad is shown
The policy ties eligibility to the display location, not the firm's home office. A firm licensed in Texas cannot serve an ad into a conversation in Florida on the strength of its Texas license. For multi-state mass tort campaigns this is the constraint that shapes the whole geo build.
United States only
Ads for legal services outside the US remain prohibited. There is no co-counsel or referral workaround: the restriction is on the ad serving into a non-US location.
Personal injury, legal claims, immigration, document preparation
The policy names these explicitly as included in what a licensed US advertiser may promote. Mass tort claim work falls under legal claims and personal injury.
Manual, case-by-case approval
Legal sits in a restricted category alongside financial services and healthcare. OpenAI says these are being rolled out gradually with approvals reviewed manually. Eligibility under the policy is not the same as an approved account, and approval is not guaranteed by meeting the written criteria.
Two reviews, then a placement filter
The advertiser is reviewed at signup, then the creative and landing page are reviewed on upload, then the placement system decides which conversations the approved ad may appear beside. An ad can clear both reviews and still barely deliver.
No political or sensitive-issue framing
Creative that meaningfully references immigration politics, abortion, race, religion, or gender identity is not permitted, even from an eligible advertiser. This catches immigration and some civil rights practices in the creative review rather than the category review.
The part nobody sells you
Ads are blocked from the conversations plaintiff firms want most
OpenAI runs a separate placement system after the ad is approved. Its job is to keep ads away from sensitive user contexts, and the policy names them: mental and personal health conversations, emotionally reliant contexts, and sensitive user journeys, alongside the obvious exclusions for self-harm, graphic violence, and political content.
Now picture the conversation a mass tort claimant actually has. Someone types that they took a medication for years and have just been diagnosed with a tumour. Someone describes a crash, an injury, a parent's death at a contaminated base. That is a personal health conversation and a sensitive user journey by any reasonable classifier reading. The better the conversation fits your case criteria, the more likely it is to sit inside the exclusion.
Contexts ineligible for ads under OpenAI's placement policy
- Mental and personal health conversations
- Sensitive user journeys
- Emotionally reliant contexts
- Suicide / self-harm
- Graphic violence
- Political content
- Debated social content
- Illicit content
This is not a targeting setting, a bid problem, or something a competent agency optimizes around. It is upstream of the auction. It is also the single strongest argument for putting the first dollar into AI search optimization instead: the organic answer has no placement filter. When a claimant describes their diagnosis and asks whether they have a case, ChatGPT still answers — it just will not carry an ad beside it. Being named in that answer is the only way to be present in the conversation that matters most.
What it costs, and why the cheap click is the wrong headline
Every agency page on this topic leads with the same comparison: a few dollars per click on ChatGPT against $50 to $100-plus for competitive personal injury terms in Google Ads. The arithmetic is real and it is close to meaningless. Legal inventory on this platform is thin, very few firms are approved, and a cheap click on an immature auction with almost no competitors tells you nothing about what a case costs.
Plaintiff firms are bought on cost per signed retainer, not cost per click. A $3 click that converts to a qualified, signed claimant at a tenth the rate of a $60 Google click is the more expensive click. Nobody on this channel has the volume yet to quote a defensible cost per retainer for a tort, including us, and you should treat a confident number from anyone as a sales figure rather than a measurement.
There is a second measurement problem specific to this work. Plaintiff intake does not close in a week. Between first contact, screening, records, and signature, the path from click to signed retainer routinely runs weeks, and the platform's conversion window is short enough that the dashboard will undercount the retainers you actually get. Judge the channel on your own intake records, matched back to session, not on what the ads platform is able to see. Our lead generation cost benchmarks and conversion rate data give you the comparison baseline from channels that have enough history to be trusted.
ChatGPT ads vs being the answer
These are different products that happen to share a surface. One buys a labelled slot beneath the answer. The other gets your firm named inside it.
| Dimension | ChatGPT Ads | AI Search Optimization (AEO / GEO) |
|---|---|---|
| What you are buying | A sponsored placement below a ChatGPT answer, labeled and visually separated from it | A citation or recommendation inside the answer itself |
| Who can see it | Logged-in adult users on the Free and Go tiers only. Plus, Pro, Business, Enterprise, and Edu are ad-free | Every tier, plus Perplexity, Gemini, Claude, Copilot, and Google AI Overviews |
| Conversation coverage | Blocked from sensitive contexts, which is where much plaintiff intake conversation lives | No placement filter. The model answers the question it is asked |
| Time to first result | Days to weeks, gated on manual advertiser approval | Months. Entity corroboration accumulates gradually |
| Cost shape | Variable cost per click, paid on every click whether or not it qualifies | Fixed program cost, no per-click charge |
| Durability | Ends when the budget ends, or when policy changes again | Compounds, but can be lost to a model update or a competitor's corroboration |
| What you control | Budget, geo, creative, landing page | Page structure, structured data, entity facts, third-party sources |
The honest recommendation for most plaintiff firms is sequence, not either-or. Build the organic position first, because it reaches every tier and every conversation and it is the asset you keep. Layer a funded ChatGPT ads test on top once your mass tort PPC program is mature enough that the test budget is genuinely incremental. Firms whose paid search is still unprofitable should fix that first; this channel will not rescue an economics problem.
State bar rules do not pause for a new channel
A sponsored placement in ChatGPT is attorney advertising. Your state's rules on outcome claims, testimonials, specialization statements, required disclaimers, and record retention apply to the ad copy and to the landing page behind it, exactly as they do to a search ad or a TV spot.
Two things make it harder here than in search. The ad unit is short, so jurisdictions that require a disclaimer or a responsible-attorney line leave you very little room to work in. And because ads serve by display jurisdiction rather than by your office location, a multi-state mass tort campaign can put one piece of creative in front of users in a dozen states with a dozen different rule sets. The compliance review has to cover every state you serve into, before the account opens, not after a grievance.
There is also an open question worth raising with your ethics counsel rather than your agency: contextual placement against a conversation in which someone has just described their injury sits closer to solicitation than keyword-matched search advertising does. No bar has ruled on it. We flag it, we document the targeting, and we keep the creative on the conservative side of the line.
How we run a ChatGPT ads test
Structured as a test with a decision at the end of it, because that is what the evidence currently supports.
Eligibility and bar review
We confirm licensure in every state you intend to serve ads into, map that to the geo build, and run the intended creative past your state bar's advertising rules before an account is opened. Jurisdiction of display, not jurisdiction of office, drives the list.
Advertiser approval
Account setup and verification, then the manual review that legal-category advertisers go through. We prepare the landing page and creative to survive the second review rather than discovering the rejection reason after the fact.
Contextual build
ChatGPT ads target conversation context, not keyword match. We translate the tort and practice-area queries that already convert for you in paid search into the context signals the platform accepts, and exclude the framings that trip the sensitive-context filter.
Attribution that outlasts the platform's conversion window
Plaintiff intake routinely runs longer than a platform attribution window measured in days, so last-click reporting inside the ads platform will undercount signed retainers. We pair platform pixels with our own session attribution and intake records so the channel is judged on retainers, not on what the dashboard happens to catch.
Read the test honestly
A monthly read on delivery, qualified lead rate, and cost per signed retainer against your existing channels, with a recommendation to scale, hold, or stop. Low delivery from placement filtering is a finding, not a failure to optimize around.
ChatGPT ads for lawyers: common questions
Can law firms advertise on ChatGPT?
Yes, in the United States. OpenAI's ad policy permits ads for legal advice, representation, and legal services when the advertiser is licensed to practice law in the jurisdiction where the ad is shown, and it names immigration, personal injury, legal claims, and document preparation as included. Legal services were added in policy version 1.5 in August 2026; before that they were disallowed, which is why a lot of published guidance still says law firms are banned. Legal services advertising outside the US remains prohibited.
Why do so many articles say ChatGPT ads are banned for law firms?
Because they were, until August 2026. Most law firm marketing pages on this topic were written between February and June 2026, when legal services sat in the disallowed list, and they have not been updated. OpenAI's own policy page also still carries a legacy summary sentence that lists legal services among disallowed categories, which contradicts the detailed Legal services section further down the same document. The detailed section is the operative one, and the changelog confirms the August 2026 change.
How much do ChatGPT ads cost for lawyers?
Reported cost per click in legal categories has been running in the low single digits, far below the $50 to $100-plus that competitive personal injury keywords cost in Google Ads. Treat those figures as early-auction pricing rather than a durable rate: legal inventory is thin, few firms are approved, and the cost per click on an immature auction tells you very little about cost per signed retainer. The number that matters for a plaintiff firm is what a retainer costs, and on this channel nobody has enough volume yet to quote one honestly.
Will my ads actually show when someone asks ChatGPT for a personal injury lawyer?
Often they will not, and this is the most important thing a plaintiff firm should understand before budgeting. OpenAI's placement policy keeps ads out of sensitive user contexts, which explicitly include mental and personal health conversations, emotionally reliant contexts, and sensitive user journeys. A claimant describing an injury, a diagnosis, or a death in the family is having exactly that kind of conversation. Eligibility to buy the ad and the ad being allowed to appear next to your best-fit conversations are two different systems, and the second one is not something an agency can optimize around.
How is ChatGPT ad targeting different from Google Ads?
Google matches an ad to a query. ChatGPT places an ad against the context of a conversation, so you are describing the situation a user is in rather than bidding on a string they typed. In practice the build starts from the same place, the queries and case types that already produce signed cases in paid search, but it is expressed as context rather than keyword match, and there is no query-level report to optimize against afterward.
Who sees ads in ChatGPT?
Logged-in adult users in the US on the Free and Go tiers. The Plus, Pro, Business, Enterprise, and Edu tiers are ad-free, and Free-tier users can opt out of ads in exchange for fewer daily messages. Your reachable audience is therefore a subset of ChatGPT's user base, skewed away from the people paying for the product.
Do state bar advertising rules apply to ChatGPT ads?
Yes. A sponsored placement is attorney advertising like any other, so your state's rules on outcome claims, testimonials, specialization statements, required disclaimers, and record retention apply to the creative and to the landing page behind it. Two things make this channel harder than search: the ad unit is short, which makes required disclosures awkward to fit, and ads serve by display jurisdiction, so a multi-state campaign can put your creative in front of users in a state whose rules you have not reviewed. We run bar review before an account is opened.
Should my firm run ChatGPT ads or invest in AI search optimization?
For most plaintiff firms, AI search optimization first. The placement filter keeps ads out of a large share of claimant conversations, the addressable audience excludes every paid tier, and the policy has been revised six times in seven months. Organic citation work reaches every tier, every engine, and every conversation, including the sensitive ones ads cannot touch. ChatGPT ads are worth a funded test for firms already running mature paid search, with budget that is genuinely incremental and a willingness to read a negative result as a result.
Is OpenAI likely to change the policy again?
It has revised the ad policy six times between March and September 2026, including reversing its position on legal services. Version 1.6 also added language reserving the right to decline ads that conflict with OpenAI's advertising principles, business interests, or competitive position, which is broad discretion held by a single gatekeeper. Build the channel as a test alongside channels you control, not as a replacement for them.
Keep going
Related reading
Mass tort signed retainers
Buy signed retainers instead of leads — pricing, medical-record standards, and case rejection terms.
What mass tort lead generation costs
Cost per lead and cost per signed case by tort, with the budget model we use to forecast campaigns.
All services
The full engagement menu — lead generation, intake, paid media, SEO, and plaintiff acquisition.
Tort-specific landing pages
Conversion-built landing pages and intake forms designed for paid traffic and TCPA consent.
ChatGPT ads for personal injury lawyers
Which PI case types clear ChatGPT's sensitive-conversation filter, context hints, licensure-matched geo targeting, and break-even CPC math.
Ready to review your next mass tort campaign?
Tell us about your firm, target cases, and intake capacity. A strategist will respond in under 5 minutes during business hours with practical next steps.
Speed-to-lead is the largest single lever in intake conversion: the gap between a 5-minute and a 30-minute callback is measured in retainers lost, not opportunities lost. The same clock is running on the torts you have not claimed yet.
Built for personal injury firms, intake teams, and mass tort dockets