To set up a ChatGPT ad campaign in June 2026, pick one of three access paths (self-serve at ads.openai.com with no minimum, Criteo, or direct), then walk the wizard through 8 setup steps: objective, budget and bid, Context hints and geo-targeting, creative, tracking, and launch. The full setup is about 4 hours of active work plus a 24 to 48 hour creative review window before live impressions start.
- Three access paths: self-serve Ads Manager (April 10, 2026), Criteo Commerce Media pipeline, or direct OpenAI sales
- Minimum spend: None as of May 5, 2026 (was $50K from April 13 to May 5; was $200K at launch)
- Targeting: Context hints (plain-language descriptions, not keywords), plus geo-targeting by state, DMA, or ZIP
- Budget and bidding: a daily or lifetime budget, plus a manual max CPC inside category-bound bid floors (e.g. $3 to $5 or $7 to $10)
- Setup time: ~4 hours active work, plus 24 to 48 hour creative review before launch
Here is what you are about to build.
ChatGPT Ads went from invite-only enterprise pilot to fully open self-serve in 86 days. OpenAI launched ad testing on February 9, 2026, dropped the minimum from $200,000 to $50,000 on April 13, opened the wizard in the wild in late April, and on May 5, 2026 opened the Ads Manager to every US advertiser and removed the minimum entirely. In the weeks since, OpenAI layered on daily and lifetime budgets, geo-targeting, and the first call-to-action buttons (the week of May 21, 2026), and conversion-optimized (cost-per-action) campaigns reach early access on June 5, 2026. The platform is no longer behind a sales-team gate.
This guide walks the entire setup, end to end, against what is actually in the dashboard today. Eight steps, ~4 hours of active work, with 5 visuals, an interactive budget estimator, and a closing analytics view that shows what your tracking is actually for. If you need background on what ChatGPT Ads are, or the dashboard wizard in detail, those companion posts go deeper. This one assumes you are ready to build.
Setting up a ChatGPT ad campaign is no longer the hard part. Setting up a ChatGPT ad campaign that pairs with a working AEO foundation is.
STEP 1: Pick your access path
~15 minThree paths into the platform as of June 2026. The right one depends on your budget, your existing ad-tech relationships, and how much hand-holding you need.
Self-serve Ads Manager (most B2B advertisers)
The self-serve Ads Manager launched April 10, 2026 to advertisers in early access, then opened to every US advertiser on May 5, 2026 with no minimum spend (was $50K from April 13). Sign up at ads.openai.com or apply at openai.com/advertisers. Our marketer's guide to the Ads Manager UI covers the dashboard structure, the 7-step submit-for-review setup, the 4-vertical eligibility constraint, and the pixel-vs-Conversions-API choice that B2B SaaS attribution depends on. The dashboard surfaces impressions, clicks, and pixel-tracked conversions in real time and is functionally similar to Google Ads. The conversion tracking pixel and a server-side Conversions API launched broadly on May 5, 2026 (10 events, 30-day window); frequency capping remains in development.
Criteo Commerce Media pipeline
Criteo was named OpenAI's first ad tech partner on March 2, 2026 and expanded the GO platform to SMBs in the US and UK on March 31. If your brand already runs Criteo campaigns, you can access ChatGPT inventory through your existing programmatic relationship. CPMs anchor on the lower end ($25 to $35) of the published range. Often the fastest path to inventory if you have a Criteo account already.
Direct OpenAI sales
The original pilot access path, still available for enterprise brands with direct OpenAI relationships. Custom creative review, bespoke reporting, managed-service overhead. CPMs anchor closer to the $60 launch rate. Most relevant for advertisers with $250,000 plus quarterly commitments and complex creative formats (carousels, native content cards).
Decision tree
STEP 2: Apply and verify your business
~20-30 minWhichever path you picked, you need to verify your business before any campaign goes live. OpenAI runs stricter advertiser verification than Google or Meta. Have these ready before you start the application.
- Tax identification number. EIN for US businesses, equivalent for international. Must match your legal business name exactly.
- Legal business name. As it appears on tax documents, not your DBA or marketing name.
- Proof of business registration. Articles of incorporation, business license, or country equivalent.
- Billing entity details. Bank routing for ACH or a corporate card sized for your planned monthly burn (no platform-imposed minimum as of May 5, 2026).
- Authorized advertiser contact. Someone whose role can sign the advertising terms.
Submit through openai.com/advertisers. Verification typically completes within 3 to 5 business days. Most rejections trace back to tax-name mismatches; double-check the legal name before you submit.
STEP 3: Choose your campaign objective
~5 min in wizardOnce verified, the Ads Manager opens to a 3-step wizard. Step 1 is Create Campaign. Step 2 is Create Ad Group and Ads. Step 3 is Review. The hierarchy mirrors Google Ads: campaigns hold ad groups, ad groups hold ads.
The Objective dropdown in Step 1 reveals three options. Reach for impression-first CPM campaigns. Clicks for CPC campaigns. Conversions, the cost-per-action objective. Reach and Clicks have been live since launch; Conversions reaches early access on June 5, 2026.
Most first campaigns select Clicks. CPC pricing makes the math intuitive (you know what each click costs), the bid floor is visible up front, and the data you collect against UTM-tagged landing pages is more actionable than impression-only Reach data while OpenAI's measurement layer is still maturing.
The Conversions objective shipped in the UI before OpenAI connected the wizard's bidder to its conversion infrastructure. The pixel itself and a server-side Conversions API launched broadly on May 5, 2026 as a self-serve beta (they track 10 events, with a 30-day attribution window), backed by the April 30, 2026 privacy policy that formalized the data-flow disclosures. OpenAI's conversion-optimized (CPA) campaigns reach early access on June 5, 2026 for accounts that set up the pixel or Conversions API by June 1 (per OpenAI's communications to advertisers, as reported by Digiday and Search Engine Land); third-party measurement remains in development. Until your account has access, optimize on Clicks and measure conversions downstream in GA4 plus your CRM (Step 7 covers this), or see our deep-dive on ChatGPT Ads conversion tracking for the full pixel + roadmap context.
STEP 4: Set your max CPC inside the category bid floor
~10 minStep 2 of the wizard surfaces a Maximum CPC bid limit field. The behavior depends on which category your ad group is in. One ad group may reject a $10 bid with the helper text "Enter a max bid between 3.00 and 5.00." Another ad group may silently accept $7.50. Same wizard, same field, different ad group, different floor.
The most parsimonious read: OpenAI sets per-category bid bands. Marketers do not compete for top-of-the-stack on bid alone; the ceiling is hard-coded by category. This is closer to a programmatic floor with quality-score-adjusted minimums than to a Google-style auction where any bid is technically valid. The cost-modeling implication is significant: at $3 to $5 in one category and $7 to $10 in another, a $50,000 budget buys very different click volumes depending on which band you are in.
See the dashboard first-look post for the side-by-side bid-floor screenshot and the Section 3 deep dive on category-bound auction mechanics. For the cost-impact framing across $50,000, see how much do ChatGPT Ads cost.
Bid floors are not a feature; they are the auction model. You compete on creative and Context hints, not on bid alone.
Estimate your monthly spend
Budget is the variable that ripples through every other step: it sets your click volume, your daily burn, and your test cadence. With no minimum spend on the self-serve dashboard as of May 5, 2026, every business now has access to the same path; budget shapes the test rather than gating it. Move the sliders below to see what your spend buys at different category bid floors and CTR assumptions.
You set this budget directly in the wizard. As of the May 2026 Ads Manager update, each new campaign takes either a daily budget (a per-day spending cap) or a lifetime budget (a total the campaign spends across its full flight). The choice is currently offered on newly created campaigns. Pair it with the max CPC from Step 4 and the estimate below to size your first test.
STEP 5: Write your Context hints
~30-45 minThis is where ChatGPT Ads diverge most from every other ad platform. Google asks for keywords. Meta asks for interests. LinkedIn asks for job titles. ChatGPT asks: describe the conversations where your ad should appear. The field is labeled "Context hints" and the helper text reads: "Context hints provide more plain-language guidance on when your ad should appear. Describe the type of queries your ad will be most relevant to, or paste a set of keywords if you have them."
How Context hints actually work under the hood
OpenAI converts your hint into vector space. Each live ChatGPT conversation is also embedded. Cosine similarity decides whether your ad fits. There is no "exact match" or "phrase match" toggle, because there is no string-matching layer. Keywords are accepted as a fallback input; they get converted to embeddings just like any other string. The framing in the helper text tells you which paradigm OpenAI thinks is canonical.
Three writing patterns that work
- Persona plus intent. "Marketing leaders comparing CRM platforms for a 50-person remote team." Names the audience and the decision stage in one phrase.
- Question patterns. "Users asking how to set up multi-channel attribution for a B2B SaaS funnel." Mirrors the natural shape of a ChatGPT prompt.
- Topic clusters with disqualifiers. "CRM evaluation, not implementation help; mid-market not enterprise; sales-led, not product-led." Tells the embedding model which adjacent conversations to skip.
Where the hints come from
The same prompt research that earns organic citations on Perplexity, ChatGPT Search, and Microsoft Copilot is the same input that targets paid ads on ChatGPT. If you already know which prompts cite you (or fail to cite you) for your category, you already have your context hints. The organic recommendation pipeline and the paid Context hints field consume the same research. Most B2B teams duplicate the work because they treat paid and organic as separate functions; they do not have to.
Layer geo-targeting on top of Context hints
Context hints decide which conversations your ad fits; geo-targeting decides where the person having them sits. As of the May 2026 update, the wizard targets by US state, DMA (designated market area), or ZIP code, set at campaign creation or edited later on a running campaign. For B2B this matters most for field-sales territories, event-driven pushes, and excluding regions you cannot service. Leave it national if your buyer is location-agnostic; the Context hints are still doing the heavy targeting.
Context hints turn prompt research into a paid-targeting primitive. The same prompt research that earns organic citations targets paid placements. One input, two channels.
Skip the 30 to 45 minutes Step 5 takes from scratch. The free Converter takes your top 3 Google Ads keywords plus 3 existing headlines and returns Context hints (each classified into 1 of 5 Patterns), 60-character ChatGPT headline rewrites, and an AI Visibility Lift content recommendation per row.
Run the free Converter →STEP 6: Build conversational ad creative
~60-90 minChatGPT users are reading, thinking, and problem-solving. They are not scanning a search results page looking for the best blue link. Your ad appears below an AI-generated answer, inside a conversation. It needs to feel like it belongs there.
The formula: empathize with the question, deliver a specific answer, introduce your solution, and offer a low-friction next step. Each piece should feel like a natural continuation of the conversation above it.
The difference in practice:
- Traditional ad copy: "Best Project Management Software. Try Free for 14 Days. Trusted by 50,000+ Teams."
- Conversational ad copy: "If you're comparing project management tools for a remote team, Acme PM was built specifically for async workflows. Teams typically cut missed deadlines by 40% in the first month. See how it handles your exact use case with a free workspace."
The second version works because it sounds like a recommendation from someone who understood the question. Early data shows ads with 60-plus words of explanatory content get 2 to 3 times higher engagement than short promotional messages. Users in ChatGPT expect detail.
Write at least 3 to 5 creative variants per intent cluster. Test different angles: feature-focused, benefit-focused, problem-focused, and competitor-comparison. You will not know what resonates until you have real data, and you will not know what data you have until you tag every variant with a unique UTM (Step 7).
One newer creative element to watch: OpenAI is running an early test of a call-to-action button on a small subset of ads (on the Free and Go tiers), the first direct-response element inside the ad unit itself. The button text is auto-selected from your creative and destination, drawn from a fixed set of "Shop now," "Book now," "Sign up," and "Learn more." You cannot hand-pick it yet, so the lever you control is the creative and landing page it is generated from. Write those so the implied next action is unmistakable.
Visualize your ad before you build it
Our free ChatGPT Ads Mockup Generator lets you type in your headline, description, and URL and see exactly how your Sponsored Recommendation would appear inside a real ChatGPT conversation. Useful for client pitches, internal buy-in, or confirming your creative reads as a recommendation rather than a sales pitch.
STEP 7: Set up UTM tracking and GA4 conversion events
~30 minMeasurement is still maturing on this platform. The OpenAI conversion pixel and a server-side Conversions API launched broadly on May 5, 2026 as a self-serve beta (they track 10 events, with a 30-day attribution window), backed by the April 30, 2026 privacy policy that formalized the data flows. Custom audiences for uploaded-list retargeting then shipped on May 14, 2026 in gated rollout. Lookalike audiences and third-party measurement integration are still in development. ChatGPT reports impressions, clicks, and the 10 pixel events; everything beyond that (longer windows, view-through, closed-deal revenue) is on you to track via UTMs and your own analytics stack. To install the pixel, see our GTM install walkthrough for the 5-step Custom HTML setup.
Build a measurement stack that survives the gap:
- UTM convention. Every ad variation needs a unique UTM string. Standardize on
?utm_source=chatgpt&utm_medium=cpc&utm_campaign={cluster}&utm_content={variant}. Document the convention in a shared doc so anyone tagging URLs uses the same fields. - GA4 conversion events. Define your conversion events (form fills, demo requests, trial signups) explicitly in GA4. Build a custom report filtered to
utm_source=chatgptto isolate performance. - CRM enrichment. Push UTM data into your CRM with the lead record. This lets you track each click through the full sales cycle and calculate actual ROI, not just click cost.
- Multi-touch attribution. ChatGPT often sits early in the buyer journey; someone researches in ChatGPT, then comes back via Google or direct. Last-click attribution will undervalue this channel every time. Use a position-based or time-decay model.
See organic and paid ChatGPT traffic in one dashboard
UTMs and GA4 events answer half the question. They tell you what each channel did separately. They do not tell you whether the new Context hint cluster also lifted organic citations, or whether the campaign creative matched the citation style of pages already ranking. For that, you need organic and paid stitched into a single time-series.
AI-Advisors stitches the AI citation feed (organic crawls plus referrals) and the UTM-tagged paid clicks into the same Overview dashboard you already use for Google Search Console, direct, and the rest of your channel mix. Now you can answer the only question that matters for the AI Visibility Lift: did the same audience find you in organic ChatGPT recommendations as in paid Sponsored Recommendations, and at what ratio?
The pre-launch period above shows a Blaze CRM advertiser already earning organic ChatGPT referrals (the AEO foundation Step 8 will validate) alongside steady Google Search and Direct traffic. April 15 is the moment Step 8 puts the Context hints from Step 5 into a live campaign. Paid ChatGPT drives the visible volume after launch, but organic ChatGPT also accelerates because the same Context hints are validated as AEO targets. See AI Analytics in detail for how the channel attribution model is structured.
STEP 8: Launch your campaign and validate organic AEO presence in parallel
~ongoingSubmit through Step 3 of the wizard. Approval typically takes 24 to 48 hours. OpenAI reviews creative against advertising policies (no health claims, no financial guarantees, no political messaging, no targeting under 18). If rejected, you receive specific feedback on what to fix; tighten the copy and resubmit.
Once approved, ads go live. Monitor impressions, clicks, and CTR daily for the first two weeks. Early data from advertisers suggests roughly 0.5% CTR on average, but this varies significantly by category. Do not panic if your first week looks slow; conversational ads take time to optimize.
The AEO parallel for every Context hint cluster
The single biggest mistake B2B teams make on launch day is treating paid as a standalone channel. It is not. Every Context hint cluster you launch should ship with a parallel AEO move. Same prompt research, two outputs.
- Per paid creative variant: ship one AEO-optimized landing page targeting the same prompt cluster.
- Per max CPC bid set: ship one direct-answer paragraph structured for retrieval (FAQ schema, 60-word direct answer, citation-friendly tone).
- Per UTM-tagged URL: ship one tracked weekly prompt run on Perplexity, ChatGPT Search, and Google AI Overviews to measure organic citation share.
- Per creative review approval: ship one BlogPosting + Person + Organization JSON-LD audit on the matching landing page.
- Per launch day: ship one IndexNow ping for any new content so Bing-stack engines (Microsoft Copilot, ChatGPT Search) crawl within minutes.
- Per daily monitoring routine: ship one weekly prompt-set re-run to track citation share trend over time.
The organic recommendation playbook covers the AEO side step-by-step. The 5 A's framework is the higher-order operating model that holds paid and organic together. Both link back to the same prompt research that drove your Context hints in Step 5.
Paid earns the placement. Organic earns the citation. Smart B2B teams ship both, because the same prompt research drives both.
How long does each step take?
About 4 hours of active work, plus a 24 to 48 hour creative review window before live impressions start.
What if you only have $1K-$5K to test, or you're not ready to spend yet?
With the May 5 minimum removal, every business with a US billing address can now test ChatGPT Ads at any budget. A $1K-$5K test reads directional signal: enough impressions and clicks to learn which Context hint clusters and creative angles resonate, even if it's not enough to settle a last-click ROAS debate. That said, many teams will want to prepare now and run campaigns later, or skip paid entirely in favor of organic visibility. All three are valid postures.
Start with organic visibility. Businesses that show up in ChatGPT's organic responses today get the most out of paid placements tomorrow. Run our free AI Visibility check to see where you stand. If ChatGPT isn't recommending you yet, learn how to get recommended by ChatGPT first. Organic presence is the multiplier that makes the paid layer work harder once you do buy.
Build your conversational creative now. Start developing ad messaging that works in a conversational format. Use our ChatGPT Ads Mockup Generator to visualize how your ads would look inside a real conversation. Test different angles, get feedback from your team, and have polished creative ready to go when you do launch.
Stand up your tracking infrastructure. UTM conventions, GA4 conversion events, CRM enrichment: all of this can be configured before you spend a dollar on ads. When you launch, you will be collecting clean data from day one instead of scrambling to set up measurement after your budget is already running.
Apply ChatGPT to your existing Google Ads work. Most teams reach for ChatGPT only after they've decided to launch a ChatGPT Ads campaign. The upstream use case is using ChatGPT as a copilot for the Google Ads work you already do today: writing responsive search ad copy, mining negative keywords from your search-terms report, and running an AI Visibility audit on top keywords to flag which ones are most exposed to displacement. The 90-minute weekly review cadence works whether or not you ever launch a ChatGPT Ads test.
The minimum is gone. The right question is no longer "can you afford it" but "is your AEO foundation ready to compound the paid signal." Build the foundation first, test small, then scale.
The Playbook walks through the citation audit, prompt research, and content production loop that produces both your AEO wins and the Context hints for paid.
Open the 5 A's Playbook →Frequently Asked Questions
#How do I get access to ChatGPT Ads in June 2026?
Three paths. The self-serve Ads Manager opened to advertisers on April 10, 2026 and was opened to all US advertisers on May 5, 2026 with no minimum spend; sign up at ads.openai.com or apply at openai.com/advertisers. The Criteo Commerce Media pipeline (announced March 2, 2026) lets brands with existing Criteo relationships access ChatGPT inventory through their account. Direct OpenAI sales remains the enterprise-premium path with custom creative review and bespoke reporting.
#What is the minimum spend for a ChatGPT ad campaign?
No minimum spend as of May 5, 2026, when OpenAI opened the self-serve Ads Manager to all US advertisers. The minimum was $50,000 from April 13 to May 5, 2026 (the original $200,000 floor ran from February through early April). Self-serve campaign creation through the Ads Manager opened on April 10, 2026, and the dashboard wizard surfaced in late April.
#Can I set a daily budget for a ChatGPT ad campaign?
Yes. As of the May 2026 Ads Manager update, each new campaign takes either a daily budget (a per-day spending cap) or a lifetime budget (a total the campaign spends across its full flight). You choose one when you create the campaign, and the option is currently offered on newly created campaigns. With no platform minimum since May 5, 2026, the budget you set is the real floor, so size it to the click volume your category bid floor implies rather than to any required threshold.
#Can I target ChatGPT ads by location?
Yes. The Ads Manager added geo-targeting in May 2026: you can target by US state, DMA (designated market area), or ZIP code, set during campaign creation or edited on a running campaign. Geo-targeting works alongside Context hints, not instead of them. Context hints match the conversation; geo-targeting limits delivery to the regions you select. For B2B it is most useful for field-sales territories, event-driven campaigns, or excluding regions you do not serve.
#Do ChatGPT ads have call-to-action buttons?
Partly, and only in an early test as of late May 2026. OpenAI began testing an optional call-to-action button on a small subset of ads (on the Free and Go tiers), the first direct-response element inside a ChatGPT ad unit. The button text is auto-selected from your creative and destination from a fixed set: Shop now, Book now, Sign up, and Learn more. Advertisers cannot manually choose the button yet, though OpenAI has said manual control may come later. Treat it as emerging rather than a standard field every campaign has.
#How long does it take to set up a ChatGPT ad campaign end-to-end?
Roughly 4 hours of active work plus a 24 to 48 hour creative review window. The 4-hour estimate breaks down as: 15 minutes to pick an access path, 20 to 30 minutes for business verification, 5 minutes inside the wizard for objective and budget, 10 minutes for max CPC, 30 to 45 minutes to write Context hints, 60 to 90 minutes for creative, and 30 minutes to wire UTMs and GA4 conversion events. From application to live ads, expect day 3 to 5 in calendar time.
#What is a Context hint and how is it different from a keyword?
A Context hint is a plain-language description of the conversations where your ad should appear, which OpenAI converts into a vector embedding and matches against the embedding of each live ChatGPT conversation. Keywords are exact-match strings; Context hints are semantic descriptions. Keywords are still accepted as a fallback input but the platform treats them as one possible source of hint text rather than the canonical primitive.
#When can I use the Conversions (cost-per-action) objective in the wizard?
Conversion-optimized (cost-per-action) campaigns reach early access on June 5, 2026, for accounts that set up the OpenAI Pixel or Conversions API by June 1; those dates come from OpenAI's communications to advertisers, as reported by Digiday and Search Engine Land, not a public OpenAI announcement. Until then the Conversions objective sits disabled while OpenAI connects its bidder to the conversion infrastructure (the conversion pixel and a server-side Conversions API launched May 5, 2026, tracking 10 events with a 30-day attribution window, backed by the April 30, 2026 privacy policy). Reach (CPM) and Clicks (CPC) remain available alongside it.
#How do I track conversions beyond what the pixel reports?
UTM parameters on every landing page URL plus GA4 conversion events plus CRM enrichment. The OpenAI conversion pixel and a server-side Conversions API launched broadly on May 5, 2026 and track 10 events with a 30-day attribution window; for anything past those 10 events or that window, lean on UTMs. Tag URLs with a clean utm_source equals chatgpt utm_medium equals cpc convention, define your conversion events in Google Analytics 4, push UTMs into your CRM with the lead record, and use multi-touch attribution rather than last-click. AI-Advisors stitches organic AI referrals, paid AI ad visits, and Google Search Console traffic into a single time-series so you can see all three channels on one chart.
#Can I move my Google Ads campaigns directly to ChatGPT Ads?
No. The platforms are fundamentally different. Google Ads matches keywords; ChatGPT Ads match conversation embeddings. Google Ads creative is short and punchy; ChatGPT Ads creative is 60-plus words and reads like a recommendation. Google Ads targets demographics and audiences; ChatGPT Ads targets via Context hints and category bid floors. Reusing Google copy verbatim produces ads that look out of place inside a conversation and get ignored.
#Should I run paid ChatGPT Ads and organic AEO at the same time?
Yes. The same prompt research that earns organic citations targets paid placements. Every Context hint cluster you launch should have a parallel AEO move: an AEO-optimized landing page targeting the same prompt cluster, a direct-answer paragraph structured for retrieval, and a tracked weekly prompt run on Perplexity, ChatGPT Search, and Google AI Overviews. Paid earns the placement. Organic earns the citation. Smart B2B teams run both because the work in one channel makes the other work harder.
Related Reading
- How to Optimize Landing Pages for ChatGPT Ads (the destination-page playbook for the campaigns you set up)
- ChatGPT Ads Just Opened to Every US Advertiser: What Marketers Should Do First
- Inside the ChatGPT Ads Dashboard: A First Look at CPC Setup, Bidding, and Context Hints
- ChatGPT Ads Are Now Self-Serve. Minimum Spend Dropped to $50k.
- How Much Do ChatGPT Ads Cost? CPM Range, Bid Floors (May 2026 Update)
- ChatGPT Advertising for Business: Is It Worth It in 2026?
- How to Get Recommended by ChatGPT: The Organic Counterpart
- How to Convert Google Ads to ChatGPT Ads in the AI-Advisors Platform
