On ChatGPT, Your Ad Needs to Talk, Not Shout

OpenAI opened up advertising on ChatGPT, and the first reaction from many advertisers was predictable: carve off a slice of the Google Ads budget and redirect it. Same digital ecosystem, same paid model, same intent-based logic. Makes sense, right?

Except intent on ChatGPT looks nothing like search. The difference isn't about format or placement. It's about how people use the platform.

ChatGPT isn't a search engine. It's a conversational partner. And the ads that work inside a conversation follow rules that paid search has never encountered. From creative to budgeting, everything needs rethinking.

As someone who works daily with paid media teams at difrnt., I'm already seeing where the first wave of ChatGPT campaigns runs into trouble: the assumption that what works on Google will work here too. It doesn't. And the reason is simpler than you'd expect.

The user isn't searching. They're exploring.

On Google, someone types "Nike running shoes size 10." The intent is clear, the decision is nearly made. The ad just needs to offer the right price and a convincing button.

On ChatGPT, the same person asks "what running shoes are good for asphalt if I have knee problems?" Follow-ups come next: "under $150?", "what's the difference between Asics and New Balance for overpronation?" It's a layered exploration, not a single-answer query.

Data from 2026 shows that sessions on AI chatbots last 3-4 times longer than traditional search sessions on average. Users don't arrive with a decision nearly made. They arrive with a curiosity they explore in real time. That fundamentally changes the context in which your ad appears.

In a conversation lasting several minutes, with questions and answers building on each other, an ad that interrupts with "Buy now!" isn't just ineffective. It's annoying. The ad needs to integrate into the flow of dialogue, not disrupt it.

No keywords. No familiar controls.

Google Ads gives you keywords, match types, search terms reports. You know exactly what someone searched for. You can optimize at the word level, exclude irrelevant terms, build audiences based on intent signals.

On ChatGPT, things are different. OpenAI doesn't provide visibility into user conversations. You don't know what someone asked before seeing your ad. You don't know if they were at the start of their research or close to a decision. You don't know whether the conversation tone was technical or casual.

This isn't just a reporting limitation. It's a fundamental shift in how you build your message. Without keywords to anchor on, the creative needs to work across multiple contexts. It has to be flexible enough to make sense regardless of where it appears in a user's conversation.

For agency teams, this shifts the center of gravity. The role of content grows significantly in the paid media process. It's no longer about writing ad copy within character limits. It's about deeply understanding how audiences talk when they're having a real conversation, not typing three words into a search bar.

Tone matters more than the offer

Imagine you're talking with a friend about vacation options. You're discussing Greece, comparing prices, evaluating which resort works for kids. And suddenly, someone breaks into the conversation shouting: "40% OFF! BOOK NOW! LIMITED OFFER!"

That's exactly what a classic paid search ad looks like inside a ChatGPT conversation.

Creative for ChatGPT needs to be consultative. It should provide relevant information, complement the discussion, not push for immediate conversion. Think about the difference between a salesperson who approaches you at the store entrance and a consultant who answers your specific question. The first makes you cross the street. The second earns your trust.

In Romania, where trust in online advertising runs below the European average according to the 2025 IAB Europe study, the consultative tone matters even more. Romanian audiences are quick to filter out aggressive ads. On a conversational platform, that filter is even more active.

Early data from ChatGPT Ads campaigns shows engagement rates 2-3 times higher for ads with an informative tone compared to direct promotional messages. In a conversation, people respond to relevance, not artificial urgency.

The budget doesn't transfer from Google Ads

The most common instinct is to cut 10-15% from the Google Ads budget and move it to ChatGPT. But these two channels don't compete for the same moment in the customer journey.

Google Ads captures people with immediate or near-immediate purchase intent. ChatGPT Ads reaches people in the exploration and evaluation phase, when they're still forming opinions and comparing options. Different funnel stages, different success KPIs.

Test budget should come from the channel that actually competes for the same outcome. If ChatGPT Ads primarily influences awareness and consideration, maybe the budget comes from display or social, not from search. And success criteria should be established before launch, not adjusted after seeing early results.

There's another factor: measurement. On Google, you have Search Terms Report, Quality Score, attribution models tested over years. On ChatGPT, the analytics tools are still in their early stages. You can't apply the same attribution models. Test budgets should be treated as an investment in learning, not as an expansion of existing campaigns.

One aspect few people discuss: your organic strategy on ChatGPT (GEO) is completely separate from ads. Paying for ads on the platform doesn't influence whether or how ChatGPT recommends you in organic responses. They're two different channels on the same platform, each needing its own strategy.

Content and paid can no longer work in silos

ChatGPT Ads isn't a classic advertising channel. It's a contextual content channel that happens to have a paid component. And that requires a real change in how teams work together.

Content and paid teams need to collaborate on ChatGPT more closely than on any other channel. Content no longer "delivers assets" to the paid team. Content becomes the paid strategy. The message isn't an ad appearing next to a result. It's a contribution to a conversation the user is having with an AI.

This demands more detailed briefs, more careful creative review, and audience understanding that goes beyond demographics and keywords. But the easy channels are already crowded. The ones that require different effort are the ones where you can build a real advantage.

Brands that treat ChatGPT Ads as just another paid search channel will spend budget without understanding why results don't look like Google's. Those who invest in conversational creative, set realistic metrics, and let the content team lead messaging strategy will have a serious edge.

Not because it's complicated. But because search behavior has changed, and advertising needs to keep up. On ChatGPT, the ad that works doesn't shout. It converses.