ChatGPT Ads UK: What OpenAI's New Format and Measurement Mean for Advertisers

OpenAI is turning ChatGPT into a place where people explore products as well as answers. On 5 October 2026 it announced a new visual ad format, a wider set of measurement partners, and brand suitability pilots with DoubleVerify and Integral Ad Science. For UK brands already buying Google and Meta, the question is practical: what are ChatGPT Ads, who sees them, and how do you test without inventing results?

This guide covers ChatGPT Ads UK from an advertiser's point of view. At BlackFire Marketing, we run paid media for UK and US brands, so the focus is sales and qualified leads, not vanity reach.

What ChatGPT Ads are (and are not)

ChatGPT Ads are sponsored units that appear below a ChatGPT response. According to OpenAI's Ads Basics help page, each ad can include:

• Advertiser name
• Favicon (logo)
• Title (headline)
• Copy (description)
• Landing page
• Image creative

Delivery is based on expected relevance and outcomes. Signals include the current conversation, the landing page, title and copy, advertiser-provided context hints, and, when personalisation is on, select signals from the user's wider ChatGPT experience. Context hints describe topics or keywords where your offer may fit. They are not exact-match keywords and do not guarantee placement in a given chat.

OpenAI says advertising does not influence the answers ChatGPT provides. Ads are labelled and kept separate from the organic reply. That is the main product claim you should test against, not take on faith in a case study you invent.

ChatGPT Ads are also not a full replacement for search or social. Inventory is thinner than Google Search. Digiday quoted Zeno Group's Shamsul Chowdhury noting roughly 2.5 billion ChatGPT prompts a day globally versus about 15 billion Google searches, with only one ad placement on ChatGPT versus multiple placements on Google. Treat that as trade-press context for why budgets need pacing, not as a BlackFire benchmark.

Who actually sees the ads

OpenAI does not show ads to users on Plus, Pro or any Business plan, or to accounts identified as under 18. Ads currently sit with Free and Go users who are eligible under OpenAI's age and plan rules.

For UK advertisers, that matters. Many heavy ChatGPT users pay for Plus or Pro and will not see your ads. Free and Go traffic can still be high intent when someone is comparing products, planning a purchase or asking how to solve a problem your product addresses. Judge the channel on downstream sales and qualified leads, the same way you would treat a new paid social or organic versus paid social test.

OpenAI reports that ChatGPT reaches 1.2 billion people each week. That is OpenAI's published platform figure, not a claim about how many of those people will see UK ads in your category.

What changed on 5 October 2026

OpenAI's 5 October announcement covered three advertiser-facing updates.

A new visual ad format in image generation

OpenAI is testing a visual ad format during image generation in ChatGPT for Free and Go users. Ads are clearly labelled and remain separate from the image being created. Testing was scheduled to begin later in October 2026 in the US with an initial group of advertisers. Digiday reported the same timetable: image ads as a US test later that month.

UK advertisers should treat the visual format as US-first until OpenAI confirms wider availability. The standard below-response unit with name, favicon, title, copy, landing page and image remains the baseline creative to prepare.

Expanded measurement and partner stack

OpenAI said it is expanding tools so advertisers can use systems they already run. Published partner areas include:

• Data and conversion plumbing: Hightouch, Tealium and LiveRamp
• Attribution and Conversions API-style partners: AppsFlyer, Triple Whale, Adjust, DV Rockerbox, Northbeam, Branch, Singular, Kochava, Airbridge and Tenjin
• Full-funnel and advanced measurement: Fospha, Measured and INCRMNTAL
• Geo incrementality (early stage): Haus, Measured and WorkMagic

OpenAI also published early partner findings. Label these as OpenAI or partner-reported, not agency results:

• DV Rockerbox: WeightWatchers' attributed CPA on ChatGPT Ads was 15.3% lower than its blended paid-search benchmark
• WorkMagic: for wellness brand Dose, 67% of incremental purchases came from net-new customers
• Triple Whale: 93% of Portland Leather's visitors from ChatGPT Ads were new

Those figures come from named partners in OpenAI's post. Use them as directional proof that some advertisers are measuring outcomes, not as a forecast for your account.

Brand suitability pilots and Negative Phrases

OpenAI is running brand suitability evaluation pilots with DoubleVerify and Integral Ad Science in controlled environments, without access to real user conversations. Qualifying advertisers can use Negative Phrases for narrower brand placement rules. Placement guardrails still aim to keep ads out of emotionally vulnerable, sensitive or otherwise unsuitable chats.

What that means for UK advertisers right now

Self-service access

As of OpenAI's Ads Manager availability list, the United Kingdom is listed as Available for self-service. The legal entity that advertises and is billed must be based in a listed country. Sign-up is at ads.openai.com. Larger advertisers can still use OpenAI's Ads team or partners where that route fits.

Category and policy caveats

OpenAI's Ads Policies treat many health, financial and legal categories more tightly outside the US. The official eligibility table (US versus All Other Countries) shows, for example, that legal services are not allowed outside the US, and that many health categories and several financial products are not allowed or are limited to select approved advertisers in other countries. Always check the live Ads Policies / eligibility table before you build creative for a regulated offer. Policies change; do not rely on a blog summary as your compliance source.

Inventory and seasonality

ChatGPT Ads remain early. Digiday noted that since ads launched in February, advertisers have pushed for better measurement, more formats and more inventory. Q4 pressure, including Black Friday planning, will compete for the same limited placement. Cap tests. Do not move core Google or Meta budgets wholesale into ChatGPT because a press release looked strong.

This channel also sits next to other AI surfaces you may already be watching, such as Google AI Mode in paid search, GEO and AEO for UK SMEs, and agentic commerce. ChatGPT Ads buy attention inside ChatGPT. They do not replace SEO or search ads that capture demand on Google.

How to test ChatGPT Ads without wasting budget

1. Confirm eligibility first. Check plan, category and country rules in Ads Manager and the Ads Policies. If your vertical is restricted outside the US, stop before you write ads.
2. Start with one offer and one landing page. Use a product or service you already measure well on Google or Meta, so you have a fair baseline.
3. Write for conversation, not keyword match. Build context hints around problems and purchase stages people discuss with ChatGPT, then keep title, copy and image aligned with the landing page.
4. Bid in line with OpenAI's guidance. For CPC campaigns, OpenAI suggests starting with a maximum bid of about $3 to $5 USD per click. Auctioning is relevance-weighted and second-price. Use Ads Manager bid-strength guidance if it is shown.
5. Choose Views, Clicks or Conversions deliberately. Impression buying suits reach tests. Click buying suits traffic tests. Conversion-optimised campaigns with click billing (oCPC) still charge for valid clicks, not for the conversion itself.
6. Instrument before you spend. Set conversion measurement in Ads Manager Beta, add UTMs to landing URLs, and connect any partner stack you already trust (for example Triple Whale, AppsFlyer or your existing CDP path via Hightouch, Tealium or LiveRamp).
7. Expect reporting lag. OpenAI says impressions, clicks and CTR refresh about every 15 minutes, while spend insights may lag 7 to 8 hours. A zero spend figure early in the day is not proof nothing accrued.
8. Judge on sales and qualified leads. Clicks alone are not enough. Use the same standard you apply in our guide to measuring marketing ROI.

Where ChatGPT Ads sit next to Google and Meta

Think in roles, not winners.

A sensible UK mix for most SMEs: keep Google and Meta as the core, then run a capped ChatGPT Ads test on one proven offer. If social search already feeds discovery into your funnel, treat ChatGPT as another discovery-to-decision surface, not a parallel brand campaign with no landing-page discipline.

Measurement checklist

Before you call a ChatGPT Ads test "done":

• Ads Manager conversions configured for the events you actually care about
• UTMs on every landing URL
• Partner attribution or incrementality path chosen (or a conscious decision to use platform reporting only for phase one)
• Brand suitability needs documented (and Negative Phrases requested if you qualify)
• Weekly review of spend, CTR, CPC or CPM, conversions, and downstream CRM or ecommerce outcomes
• Written decision rule: continue, pause or scale, based on cost per sale or cost per qualified lead, not on impressions alone

OpenAI's partner examples (WeightWatchers via DV Rockerbox, Dose via WorkMagic, Portland Leather via Triple Whale) show that serious measurement is possible. Copy the discipline, not the percentages.

FAQ

Yes for self-service Ads Manager, according to OpenAI's availability list: the United Kingdom is marked Available, and the billed legal entity must be based in a listed country. Confirm live status on OpenAI's help page before you plan spend, because country lists change.

OpenAI does not show ads to Plus, Pro or Business users, or to under-18 accounts. Eligible Free and Go users may see ads below responses when the conversation context is suitable under OpenAI's placement rules.

OpenAI is testing a visual format during image generation for Free and Go users, initially in the US later in October 2026. Ads are labelled and kept separate from the generated image. UK advertisers should assume the standard below-response unit until OpenAI opens the visual format more widely.

Advertisers set a maximum CPM for Reach or a maximum CPC for Clicks. OpenAI recommends starting around $3 to $5 USD CPC for click campaigns. Selection uses a relevance-weighted, second-price auction. Conversion-optimised campaigns with click billing still charge for valid clicks.

Often with tighter limits than in the US. OpenAI's eligibility table restricts or blocks many health, financial and legal categories outside the US. Check the current Ads Policies and eligibility table in OpenAI Help before building creatives.

Next step

If you want a capped ChatGPT Ads UK test alongside Google and Meta, BlackFire's paid media team can help you set eligibility, tracking and a decision rule before money goes out. Get in touch with the offer you want to test and the cost-per-sale or cost-per-qualified-lead target you already use.