Conversational AI Is Creating a New Media Channel. It Will Take More Than One Discipline to Build It.

By Amanda Wallingford, Programmatic Director, and Sophie Ryland, Associate Director, SEM 

For years, digital media has been organized around two complementary behaviors. Search captures what people explicitly ask for. Programmatic builds relevance through audiences, content, creative, and repeated exposure. Conversational AI is starting to bring those behaviors together, and it’s happening faster than most channels take to mature. 

Consumers are no longer just typing a few words into a search bar. On platforms like ChatGPT, which 44% of U.S. adults surveyed say they use, they’re asking detailed questions, comparing options, and requesting personalized recommendations through ongoing conversations. That creates a new kind of discovery environment, shaped by intent that’s broader and more nuanced than a single keyword. It also raises an organizational question: who should own it? 

Why Conversational AI Can’t Be Built in Silos

As we see it, conversational AI shouldn’t sit entirely within Search or Programmatic. It should be built by both. Search team’s understanding of query evolution and search intent while the programmatic team’s knowledge of ad activation and brand suitability allows for Conversational AI to play to both teams’ strengths. Together, they can treat Conversational AI as more than a new ad format; it’s a new meeting point between intent and discovery. Rather than replacing either channel, Conversational AI adds another layer into the media mix, supporting exploration and comparison while Search captures active demand, and Programmatic builds and reinforces relevance. 

The pace of change here is unlike a traditional media launch. Building an entire channel on evolving requirements, ad formats, and attribution requires steadfast flexibility. Previously siloed workflows now have intentional pivot-points through Conversational AI, which has already elevated search and programmatic strategies to be able to act on curiosity when it arises. 

Pressure Test: Our First Self-Service Campaigns

We’re currently running our first self-service test through StackAdapt for an August launch, with a second self-service test planned directly through OpenAI in Q3. Running both paths lets us compare platform capabilities, targeting, and reporting as the ecosystem matures, rather than betting everything on one activation route before we understand the tradeoffs. Right now, ads run only against free-tier users 18 and older, which still covers the vast majority of the platform’s weekly base. Targeting is limited to geography and phrase matching layered on top of intent, with no audience data available yet. Campaigns can be bought on a CPM or CPC basis, with CPC generally better aligned to engagement objectives and CPM supporting reach. 

How Brands Should Show Up and How to Measure It

Consumer journeys rarely follow the tidy funnel we draw on a slide, and Conversational AI fits that reality well. Someone might start with a broad question, compare options, step away, and return later with something more specific. The opportunity for brands isn’t to interrupt every conversation or force a conversion. It’s to show up with a useful next step when the context genuinely supports it. That’s part of why we’re already seeing interest from a wide range of categories — retail, tech, and travel among them — all drawn to a platform that can aggregate and organize information quickly for a user actively comparing choices. 

That opportunity requires restraint as much as innovation. Marketers should resist two extremes: waiting until every question about the channel has been answered, or scaling aggressively before it has established real benchmarks. A better approach is structured experimentation, starting with a defined use case and clear stakeholder alignment across Search, Programmatic, Analytics, and Creative. Initial measurement should extend beyond a single efficiency metric. Clicks and site activity matter, but so does visit quality, which contextual themes perform, and how this channel complements search and programmatic rather than competing with them.  Its value should be evaluated not only as a stand alone placement, but also by how well it strengthens existing Search and Programmatic efforts. 

Conversational AI is not simply Search with longer queries, or Programmatic in a new placement. It’s a meeting point between intent and discovery, and building it well requires both disciplines working together, not in sequence.