ChatGPT Ads' Platypus Problem

OpenAI has started testing ads and most people are treating it like a search product. That's the wrong frame, and it's going to lead to the wrong expectations, the wrong measurement, and ultimately the wrong conclusions about whether it works.

The Setup: What We Know

Ads will appear after the AI completes its organic response. OpenAI has been explicit: Advertisers won't influence what the AI says. There’s no pay-to-play inside the actual answer.

That promise is going to be tested immediately. Skepticism is already public, with Anthropic questioning whether ads can work in chat interfaces at all. The concern is valid: When ads sit right next to recommendations, it's hard to convince people that money isn't influencing the advice. The first brand scandal—real or perceived—could crater trust in the entire model.

But monetization is coming regardless. The question isn't if this traffic gets monetized, it's which form factor wins. For ads to work here, they need to complement the organic response, not compete with it. They need to be genuinely useful. There are plenty of scenarios where the AI response educates but doesn't complete the journey. That gap is where advertising can add value.

The catch is in the details. And the biggest detail everyone's missing: This isn't actually a search product.

The Platypus: It Looks Like Search, But It's Not

AI optimization—GEO, AEO, whatever acronym you prefer—is being treated like an extension of SEO. Do your usual optimization, add some AI-specific tactics, and you’re set. That framing misses what's really happening.

When you optimize for AI responses, you're not trying to rank in a list of links. You're trying to influence how an artificial intelligence thinks about your brand. You're trying to make it believe you're the right answer for a given query.

Sound familiar? That's brand marketing.

Brand marketers don’t push for immediate conversion. They build associations, establish credibility, and create preference over time. That groundwork is what makes lower-funnel tactics work later.

AI optimization is a platypus. It looks like SEO on the surface —you're working with content and trying to "show up"—but it behaves like brand advertising. You're influencing synthesized judgment, not just gaming link placement.

Why Treating These Like “Search Ads” Is a Big Mistake

If the organic side is fundamentally brand-building work, the ads need to be understood the same way.

AI is used for discovery, consideration, and comparison. Yes, people can move from discovery to purchase in a single session, but that usually takes time and multiple turns. The behavior looks nothing like traditional search, where someone types "running shoes," clicks an ad, and expects to buy immediately.

A more realistic scenario looks like this:

A user asks a question, the AI provides a thoughtful response and mentions a few brands. One of those brands appears in an ad right after the response.

That brand now has an advantage—if it uses it correctly. The ad shouldn't push "Buy Now." It should extend the conversation: Apply this to your situation. Build a personalized plan. See how this works in practice. The goal isn’t conversion. It’s continuation.

And here's the tricky part: the user might not click. They’ll keep chatting. They’ll leave. Then hours or days later, they’ll visit the site and convert.

That's where the value sits. Not in direct response but in assisted conversions and view-through attribution. The same measurement frameworks used in programmatic display today—brand lift, consideration, contribution to the journey—adapted for AI environments.

If this is measured like search, with last-click attribution and direct ROAS expectations, it will look like a failure even when it’s doing meaningful work.

The Budget Problem

If this isn't search, where does the budget come from?

Pulling from search sets the wrong expectations. Teams will expect search performance—high CTR, immediate conversions and tight attribution. When they don't see it, they'll pull spend.

But moving it into brand or programmatic display creates its own friction. Different owners and KPIs and even different politics. Creating a new line item is even harder. The result is that every test becomes a referendum on whether "AI ads work," rather than a learning exercise about how they work differently.

When Your Brand Gets Mentioned

AI rarely gives a single answer. Ask for a CRM recommendation or running shoe advice and you'll get three or four options. When your brand shows up, it’s almost always alongside competitors.

That moment matters. The AI has effectively told the user that all of you are viable options. This is where the ad earns its keep—even if (especially if) you were already mentioned organically. You're not paying for visibility you would have gotten anyway. You're paying for differentiation at the exact moment the user is weighing options.

The ad gives you a chance to break out of the pack. Maybe it's a specific offer. Maybe it's a next step that feels more actionable than what the other brands can offer. Maybe it's reinforcement at the exact moment the user is deciding.

It’s also likely that ads will primarily appear for brands already mentioned in the response. Will the platform even show your ad if you weren't mentioned in the response? Probably not. It would be odd for the AI to think a brand is relevant enough to advertise but not relevant enough to mention. So the more realistic scenario is: ads will predominantly show up for brands that were already in the organic response.

That's not a problem—it's the point. But it does mean incrementality becomes critical. How do you know the ad drove the conversion versus the organic mention? You need frameworks to measure overlap, view-through impact, and assisted conversions. This is where things get sophisticated fast.

Organic and Paid Are Inextricably Linked

Unlike traditional search, organic and paid are tightly linked in AI.

The AI response always comes first. Always. The ad only works if it complements what the AI just said. If you're not showing up in the organic response, your ad has almost no chance of being relevant or effective.

There’s no realistic version of winning through ads alone. If the AI names three competitors and not you, there’s no ad copy that fixes that gap.

This flips the old search model on its head. You can’t buy your way to relevance. The organic response sets the stage. Paid media can only build on it.

Honestly, if brands had to choose, the choice is easy: Prioritize organic visibility. Because if you don't exist in the response, the ads won’t save you.

What Needs to Happen

For this ecosystem to work for them, advertisers need to:

1. Prioritize organic visibility. Before you spend a dollar on ads, make sure you're showing up in AI responses. Product feeds, structured data, owned content, brand signals—all of it matters. This is the foundation on which everything else is built.
2. Rethink measurement. Last-click attribution won’t tell the story. View-through tracking, brand lift , assisted conversion models are essential. Treat this like upper-mid funnel, not bottom funnel.
3. Rethink creative. Ads should extend the AI response, not replace it. "Learn more" is table stakes. "Build your plan" or "See how this applies to you" gives the user a reason to engage beyond the generic response.
4. Rethink content strategy. If organic visibility is brand-building work, content stops being just an SEO function and becomes a core brand asset.
5. Rethink budget sourcing. This probably shouldn't come from search. It might come from brand, programmatic, or a new line item. But wherever it comes from, the expectations must match reality. This is about influence and consideration, not immediate conversions.

The Bigger Picture

We often confuse the act of searching with the platforms where it happens. People search on Amazon, YouTube, social, and app stores, each with different intent and outcomes.

AI search is another context. You're not scanning links. You're in a conversation. The AI synthesizes information, makes recommendations, and shapes understanding in real-time.

That's closer to a consultation than a "search." And if the experience is consultative, the advertising needs to feel like a natural next step, not an interruption.

The Bottom Line

Post-response ads are coming. Monetization is happening. But this won’t behave or perform like search.

It's a platypus: It looks like one thing, but it plays by different rules. Until advertisers accept that, we're going to see a lot of failed tests and premature conclusions.

The opportunity is real. But it requires rethinking how we measure, how we create, and how we think about the role of paid media in AI environments. That's a bigger lift than most teams are ready for.

But the ones who get it right early will have an advantage. Because this isn't about gaming an algorithm. It's about shaping how AI thinks about your brand. And it starts with organic. Always.