Inside Conversational AI: What Early Advertisers Are Learning
Key Takeaways for Success from Wpromote’s Webinar with OpenAI and Kargo
OpenAI launched its ads business in February. A few months later, paid advertising in conversational AI has moved from a space with uncertain tactics to a channel that demands its own strategy, one built around solving a real problem for the user.
That was the main argument made during Wpromote’s webinar, “Conversational AI: What Early Advertisers Are Learning.” Panelists Whitney Muldowney (OpenAI) and Alena Morris (Kargo) joined Casey Geiger (Wpromote) and Christine Schrader (Wpromote) to discuss what that looks like in practice, the strategies brands should be taking, and why testing now matters for success later.
Here are their key takeaways for success:
IT’S NOT SEARCH 2.0
“That’s the biggest misconception most brands bring to the space,” opened Casey Geiger, setting the stage for why advertising in AI deserves its own strategy. Brands are still bringing over their keyword playbook and expecting it to work the same way in LLMs. It doesn’t.
The search query compresses intent into a few words. “In ChatGPT, we’re seeing our users bring objectives, constraints, context, and trade-offs. And we’re finding them over several turns within the same conversation,” explained Whitney Muldowney.
Casey Geiger gave a retail example. A search box gets “black tie wedding attire.” Conversational AI gets “What do I wear to a black tie wedding?” Follow-up questions might explore the venue, the season, and the couple’s own style. Same topic, but a vastly different amount of context to work with.
By the end of the exchange, what might have started as a question about black tie wedding attire is now an intentional search for a summer-weight navy tuxedo.
THE FUNNEL HAS COLLAPSED
Before AI, someone planning a summer trip to Europe might scroll Instagram, check reviews on Reddit, watch YouTube videos, and then, finally, head to Google.
Today, when someone starts an AI chat with “I want to plan a summer trip to Europe with my kids,” the conversation can take a number of twists and turns as it works through budget, preferences, and destinations, and ends with booking a hotel, all from inside the platform. That’s not a metaphor for the funnel collapsing. It is the funnel collapsing.
Awareness, consideration, and conversion are all happening in the same space. The context of the entire journey directly influences what ad is shown.
HOW PAID ADVERTISING WORKS IN CONVERSATIONAL AI
An ad in conversational AI shows up at the end of the conversation, only after the decision’s been made and the user’s ready to take action.
As Alena Morris from Kargo put it, “I love to host dinner parties. For a recent one, I wasn’t sure what to make for dessert, so I asked ChatGPT. It gave me ideas and a shopping list, and only then did a native ad for a local grocery store show up.”
“What we’re seeing through a lot of our initial testing and the campaigns is that when ads feel super native in the context of the conversation, clickthrough rates are doubling,” she added.
Whitney Muldowney confirmed as much from the OpenAI side. Paying for an ad doesn’t change what ChatGPT says to a user. Paid ads can’t shape, rank, or touch what OpenAI’s platform actually surfaces. Paying OpenAI buys advertisers a spot next to the answer that already fits the customer’s decision.
THE FRAMEWORK FOR SUCCESS
Regardless of who was speaking, the panel’s answer for how to succeed was resoundingly clear. Solve the problem the customer is surfacing.
“The brands that win here are the ones with a clearly defined problem they’re solving for the customer,” stated Casey Geiger. Whitney Muldowney framed the payoff as meeting someone “in that key decision layer” with something genuinely useful.
Another way to think about it: paid advertising in conversational AI acts like customer service. Alena’s grocery store ad didn’t interrupt her dinner party planning, it helped her finish it. That’s putting the customer first, the oldest idea in marketing, just showing up somewhere new.
THE THREE LEVERS FOR PAID ADS
Knowing what problem to solve isn’t enough on its own. Webinar attendees also got exclusive insight into the three levers that actually help get a paid ad surfaced. Significantly, only one of them lives inside the ad:
- Ad Copy: The ad’s title and description tell AI whether your ad answer matches what it’s about to recommend.
In one OpenAI case study, an advertiser scaled their creative from 300 to 20,000 variations. Each one was written for a specific situation their customer might be interested in. As a result, their CTR jumped nearly 50% with most of the traffic coming from new users.
- Web URL: Web crawlers go onto your site, scrape content that goes into the ad auction, and serve it right to the user.
The same goes for other content you produce across the web. A consistent brand with a clear value proposition determines how and when you get surfaced.
- Context: Signals from inside the conversation match the right ad to the right moment.
Keep your product feeds and targeting data clean, as that’s what the algorithm reads at the moment.
Together, the three levers point to one of the biggest tells about strategy in this space. Success in conversational AI means treating paid and earned as one connected play.
If any of this sounds overwhelming, it doesn’t have to be. There are a few smaller steps to getting traction in paid AI first.
WHAT TO DO FIRST
Asked for the one thing a brand should do to get started, each panelist had a specific answer.
Whitney: rewrite your existing search terms as customer problems, and get conversion tracking in place now, before you need it.
Alena: audit where your brand already shows up organically in chat before you spend a dollar on ads. If AI is recommending you, that’s a head start. If it skips you, that’s the gap to close first.
Casey: get clear on the actual questions your customers are asking, then decide what your own brand’s measurement standard needs to be. Your appetite for proof will look different from anyone else’s.
Each piece of advice goes back to the same fundamental framework: knowing your customer, and knowing how your brand shows up when they ask about a problem you can help them solve.
WHY THE CLOCK IS ALREADY RUNNING
The panel felt strongly that brands should treat this with urgency, because this is where customers are moving.
Gartner projects traditional search volume will drop 25% in 2026. More than half of consumers surveyed by Orbit Media now start a search by opening an AI app instead of a search engine. ChatGPT passed 1 billion active users in June 2026.
Of course, the space is not fully built yet. Attribution windows are still being established, measurement is playing catch-up, and platforms are still experimenting with their ad models.
Like any emerging channel, a brand needs to approach the space with a test and learn mindset. Starting now gives brands the opportunity to help shape what “best practices” look like from the beginning “2026 is a must-have conversation, and 2027 is a must-have strategy for everyone. This isn’t a passing channel. It’s not a trend. Ads and AI are going to be table stakes in the same way that paid search and paid social are right now. Waiting doesn’t remove the risk. It just hands the head start to a brand that moves first,” summarized Casey Geiger.
FINAL TAKEAWAY
Conversational AI belongs in the media mix as its own dedicated strategy. One that starts with knowing how your brand serves your customer.
Success in AI advertising comes down to a unified approach. Paid ads are an increasing part of a holistic strategy, the other portions being earned and technical approaches. Treat your AI marketing as a single, connected play, and you’re poised for success.
You can catch all the insights and recommendations by watching the webinar on demand. If you are ready to talk about AI strategies, you can reach us here.
This is the first post in a series on what early advertisers are learning in conversational AI. More on measurement, scaling, and brand safety ahead.
