OpenAI Just Switched ChatGPT Ads to Pay-Per-Click. The Playbook for Marketing Agencies.

On April 21-22, 2026, OpenAI quietly switched its ChatGPT advertising from cost-per-mille (CPM) to cost-per-click (CPC) pricing. For marketing agencies, this isn’t a footnote buried in ad-ops news. It’s a structural shift in how AI search will monetise, and it tells you exactly what’s coming next across Perplexity, Gemini, and Claude.

Here is what happened, why it happened, and the five moves every agency should make this week.

What actually happened

Per Digiday, OpenAI turned on cost-per-click pricing for ChatGPT ads on April 21-22, 2026. Advertisers can now set bids between $3 and $5 per click. The minimum spend was cut from $250,000 to $50,000, a fivefold reduction.

The timeline, for context. Per OpenAI’s own announcement, ads launched inside ChatGPT on February 9, 2026 at a $60 CPM with a $250,000 minimum spend. Early advertisers included Target, Ford, Adobe, Mrs. Meyer’s, and Expedia. Per The Next Web, the CPM eroded from $60 to approximately $25 within ten weeks, which is a collapse not a decline. The shift to CPC was a response to that.

Other pieces of the rollout. Per PYMNTS, OpenAI opened its self-serve ads manager globally on April 15. Per Quartz, a conversion tracking pixel is now live, supporting events including lead creation, order creation, and subscription starts — which is the full-funnel instrumentation you need to justify ad spend. Per Skift, OpenAI has partnered with StackAdapt for programmatic placement.

The revenue arc tells the rest of the story. Per Digiday, the pilot topped $100 million in annualized revenue within six to eight weeks. OpenAI projects $2.5 billion in ad revenue in 2026, scaling to $11 billion in 2027 and $100 billion by 2030. Those are not aspirational numbers from a pitch deck. They are targets being quoted to shareholders.

Why CPM failed (and why CPC will too, eventually)

CPM-based pricing works when inventory is scarce and demand is consistent. OpenAI had neither. Ads appeared inconsistently across prompt types, advertisers could not forecast impressions, and $60 CPM on that level of unpredictability felt like a tax. CPC shifts the risk to per-click performance and drops the price of entry, which will broaden the advertiser base quickly.

The structural lesson is what matters for agencies, not the specific pricing model. OpenAI is testing, iterating, and growing ad revenue as fast as the market will absorb it. Per The Information, CPC is not an endpoint — it’s the next test. Expect more formats (sponsored recommendation cards, conversation sidebars, conditional sponsorship based on prompt category), more targeting options, and more pressure on organic visibility as paid inventory expands.

Four things this changes for marketing agencies

The paid/organic split comes to AI search

This is the pattern every marketing agency should have on a whiteboard this week. Per OpenAI’s stated approach, ads appear as clearly labelled “Sponsored Recommendations” placed beneath or within structured recommendation cards. Only Free and Go-tier users ($8/mo) see ads. Plus ($20), Pro ($200), Business, Enterprise, and Education tiers remain ad-free.

That last detail matters. ChatGPT Pro and Business users, who are disproportionately decision-makers, never see the ads. Those buyers see only organic recommendations. For a B2B agency, that means the highest-intent, highest-revenue prospects are specifically the audience where earning organic citations — not bidding for ads — is the only path to visibility.

The parallel to Google Search is direct. In 2004, the “advertising-free organic results” argument was novel. By 2010 it was just how the web worked. The transition window between those two states was roughly five years, and the agencies who moved first on organic SEO owned a decade of compounding advantage. The AEO version of that transition window is happening now.

What every agency should do this week

• Run a CPC exposure audit on every client this week.

  Which of your clients operate in categories where competitors are likely to bid? High-CAC verticals first: healthcare, legal, insurance, B2B SaaS, home services. These are where ChatGPT ads will show up first and hurt most.
• Baseline each client’s organic AEO across all five AI engines.

  ChatGPT, Perplexity, Gemini, Claude, Google AI Overviews. Record citation rate, average position, sentiment per engine. You cannot prove organic AEO value in 60 days without a clean baseline today.
• Prepare the paid-plus-organic messaging for the next QBR.

  Every client will hear about ChatGPT ads within 30 days. Have a position before they ask. The position is not “don’t run ads.” The position is “we will run them selectively and we are protecting your organic share in parallel.”
• Identify the three to five clients most likely to be squeezed.

  Clients whose CAC is already high, whose competitors have ad budget, and whose organic AI visibility is weak. Move those clients to an AEO-forward plan within the next two weeks. These are the retainer saves of Q3.
• Get organic AEO monitoring in place before the next engine ships ads.

  Perplexity and Gemini are six to twelve months behind. The agencies that have the baseline in place before those launches will close the largest Q4 retainers of 2026.

The organic moat — what agencies with AEO work already done are seeing

There’s a defensible answer to “should we run ChatGPT ads?” and it looks like this. Paid and organic are not substitutes, they are complements. Agencies that have done earned-mention work, structured their clients’ content for citation, and built multi-engine monitoring are the ones whose clients don’t bleed when paid ads arrive. That is not GenPicked marketing copy. That is how paid/organic splits have always worked in every search channel since 2004.

The GenPicked view, with that caveat: the agencies closing new retainers this quarter will be the ones who can walk into a client meeting next Tuesday and produce a baseline report showing citation rate across all five engines, gaps by query, and the three-move plan to close the biggest gaps. The agencies losing retainers this quarter will be the ones who heard about ChatGPT CPC from their client instead of the other way around.