The advertising, marketing, and media industries arrived in Miami for the fourth annual POSSIBLE conference last week carrying a familiar set of unresolved tensions: AI tools that promise transformation but routinely underwhelm; measurement frameworks that still can’t account for where attention actually goes; and a supply chain under pressure from holding companies determined to compress it.
Despite these issues, or, perhaps because of them, POSSIBLE 2026 drew more than 7,500 attendees — up from 5,400 last year — and organizers marked the moment by announcing the conference’s first international expansion, a 2027 edition in Lisbon, Portugal.
“Following the enormous success and continued growth of POSSIBLE Miami, we’re thrilled to share that POSSIBLE is coming to Europe,” said Christian Muche, global president and co-founder of the event.
For attendees like David Berkowitz, chief community officer at Marketecture Media, the milestone confirmed what the industry already sensed. “It’s cemented itself as a must-attend ad industry event,” Berkowitz says. “POSSIBLE isn’t some consolation prize for those who can’t get to Cannes but an important part of the ad industry’s calendar in its own right.”
Kevin Wassong, CEO and founder of mktg.ai, was struck by MMA CEO Greg Stuart’s opening comments, which served as a provocation: “the industry has to get back to creative.” Wassong, whose company bills itself as the “Bloomberg Terminal for marketing,” called the gathering one of the strongest POSSIBLE events in recent years, and said Stuart’s framing set the tone for everything that followed.
From there, Wassong identified three themes that ran through the week. AI is reducing the labor of reporting, but the real opportunity is closing the gap between insight and action while campaigns are still live — analysis after spend is complete is hindsight, not intelligence.
Second: AI should elevate analysts into strategists rather than replace them, but most organizations still stop at diagnosis.
Third: Enterprise adoption remains the biggest barrier, with too many companies layering AI onto fragmented infrastructure. “If the underlying marketing infrastructure is disconnected, delayed, inconsistent and reactive, AI doesn’t solve the problem,” Wassong says. “It simply scales inefficiency.”
Wassong’s skepticism extended across the conference floor. Crystal Foote, founder and CEO of Digital Culture Group, observed that the industry is in its “year of mobile” moment for AI — everyone talking about it, many brands still struggling to identify what’s genuinely useful versus what amounts to a more sophisticated dashboard.
“The gap is not ‘interest,’” Foote says. “It is ‘application.’ Marketers need AI that helps them make better decisions before campaigns launch, not just summarize data, automate tasks, or repackage existing segments.” Chris Grosso, CEO of Intersection, put it more directly: “While AI’s efficiencies are overstated, its impact on creativity is underestimated.”
Grosso flagged ad fraud as an underappreciated variable in the AI conversation. With roughly half of web traffic now driven by non-human sources, he says marketers should be reconsidering channels like premium video, podcasts, and out-of-home that are largely fraud-resistant — a structural reality that shapes where AI-driven optimization can actually deliver clean signal.
The measurement conversation was equally pointed. Nielsen set the tone on Day One when Nichole Henderson launched Predictive Sales Lift inside Nielsen One Ads and disclosed that brands are not measuring lower-funnel metrics for 90 percent of campaigns running through the platform.
Damian McKenna, co-founder of CTV performance platform Emodo, drew the implication plainly: when the conference opens with the legacy measurement company acknowledging that 90 percent of CTV is still measured on the wrong metrics, the debate over whether CTV is a performance channel is over. The question now is who closes the gap.
Jon Schulz, CMO of Viant Technology, connected that gap to Viant’s TVision acquisition. “Attention is the missing variable that has been hiding in plain sight,” Schulz says. “If someone isn’t watching your ad, it didn’t happen.” Viant’s position is that room presence, co-viewing data, and second-by-second eyes-on-screen signals shouldn’t just inform a campaign report — they should drive the buy.
Meanwhile, Vikrant Mathur, co-founder of Future Today, sees a broader shift: “Advertisers are no longer satisfied with scale alone. They want real-time insight into where their messages appear, how audiences engage and, ultimately, what business results are being driven.”
Conversations around agentic AI generated the week’s most substantive debate. Spencer Morris, EVP of data operations and data products at iHeartMedia and Triton Digital, pointed to Omnicom’s earnings call — held during the conference week — as the clearest signal that agent-to-agent buying has moved from theoretical to operational.
Morris pointed to Omnicom CTO Paolo Yuvienco’s confirmation that the holding company is executing agentic buys for clients using the AdCP protocol, noting as well that CEO John Wren’s stated goal is to compress the supply chain in order to put more working media dollars in front of publishers directly.
Yet McKenna noted that MiQ told AdExchanger on the ground that most tools marketed as agentic remain workflow automation, and that truly autonomous buying is still further off than the stage conversations implied. The buy side, McKenna says, has stopped letting vendors blur that line.
Complicating matters, perhaps is what Morris identified as an infrastructure consequence: as agent-to-agent buying matures, identity signal quality becomes the differentiator that separates smart decisions from wasteful ones. Omnicom’s own language — that Acxiom’s identity infrastructure is “exponentially powering” its agentic buys — made the point explicitly. Grosso arrived at a similar conclusion from a different, more optimistic angle: the data cleanup process required to make agentic buying work may itself unlock significant value, independent of the AI layer on top.
McKenna flagged one underreported development from the week: Meta’s move to extend its audience demand into third-party CTV inventory. “Meta’s playbook is performance density,” McKenna says. “If it brings that signal density to third-party CTV, it’ll leave the open web an even smaller window to prove it can compete on outcomes.” He observed that offstage CMOs were asking a question the main stage mostly avoided: whether the next dollar in CTV actually beats the marginal dollar in search or social. New ad formats — pause ads, shoppable units, interactive overlays — didn’t answer it. “The next dollar of performance will come from making the existing impression smarter,” McKenna says, “not from inventing a new one.”
In a counterpoint to the week’s automation anxiety, Max Lazerwitz, strategy director at Meet The People, pointed out that the POSSIBLE AI Verse sessions consistently reinforced that AI is enabling teams to do more, not eliminating them. “The more AI accelerates work, the more valuable it becomes to slow down and actually meet the people behind it,” Lazerwitz says. “In a moment where everything is being automated, the real differentiator is still human: connection, community, and trust.”
The mix of people and content at POSSIBLE 2026 said a lot about the connected states of retail media, AI, creativity, and performance, as all shared the same stage, alongside unexpected cultural curveballs like NHL standout Matthew Tkachuk, noted Sam DiGennaro, founder/CEO of DiGennaro Communications.
“It’s a clear signal that the industry conversation is no longer siloed. It’s converging.” DiGennaro said. “In just four short years, POSSIBLE 2026 feels like it’s fully come into its own; bigger in scale, broader in ambition, and more reflective of where marketing is actually headed than just about any other industry tentpole.”
