CES 2026: Why agencies are finally treating AI as software, not slideware

“There was a time when these were more slideware than they were actual software,” he says. “That’s changed. These are viable platforms that are adding value to the agency and adding value to the client.”

Across the holding companies, the pattern was unmistakable. WPP pushed further on Open and its Open Pro self-serve layer. Omnicom unveiled the latest iteration of Omni, now badged Omni+. Havas continued to bolster Converged as an end-to-end data, audience and content platform. Stagwell showcased The Machine. And Monks highlighted the steady evolution of products such as Monks.Flow.

On the surface, these platforms differ in branding, architecture and emphasis. Underneath, Pattisall argues, they are converging rapidly around the same idea: AI as infrastructure rather than embellishment.

“These platform plays represent a practical application of AI,” he explains. “Machine learning defining audiences and mining insights, generative AI producing content and, now, agentic layers sitting on top to execute workflows. From media planning to briefing to production, it’s about speed and coordination.”

The significance, however, goes beyond workflow. What Pattisall sees emerging is a structural shift in what agencies sell.

“These platforms become the predicate for transforming how agencies deliver marketing services – and how they get compensated,” he says. “It’s not just people making ads any more. It’s people using software to make ads.”

That distinction underpins the industry’s gradual move from pure services towards hybrid SaaS-and-services models. Some groups are positioning platforms as incremental value layered on to existing relationships. Others, Stagwell most visibly, are exploring more explicit SaaS revenue plays.

Even so, Pattisall is cautious about overstating differentiation. “The trend isn’t clean,” he says. “Most of these platforms have multiple go-to-market models.”

Where differences do exist, they tend to centre on data strategy. Omnicom, strengthened by acquisitions such as Acxiom and Flywheel, holds deeper proprietary data than most. Publicis and Dentsu have leaned harder into digital identifiers, while others assemble federated third-party stacks. Meaningful distinctions, yes, but not yet decisive.

“Any agency will tell you no one has what they have,” Pattisall says. “The truth is, they all have what they all have. The competitive advantage is still somewhat in development.”

That doesn’t mean agencies are misreading the moment. Pattisall believes the industry has broadly made the right strategic bet by investing in systems designed to create future advantage, even if that advantage has yet to fully crystallize.

What concerns him more is how value is currently framed. AI, he argues, has been sold upwards largely as an efficiency story.

“That’s one answer,” he says. “But it’s not the only answer. The real value is better experiences, better content and better marketing. That’s where growth comes from.”

For now, efficiency dominates because it is easier to count. Headcount reductions and productivity gains show up quickly on balance sheets. Growth takes longer – and demands that agencies help clients re-argue the case for marketing as a profit center rather than a cost line.

This perspective also explains Pattisall’s impatience with nostalgia. He has little time for romanticizing a supposed golden age of creativity.

“The industry should not look back longingly to a time that’s gone. In a fragmented environment where the audiences are both humans and machines, influence is less about a single creative idea and more about a systematic approach to marketing.”

Creativity doesn’t disappear in this model; it is resituated inside systems. Coordination, data and execution matter as much as inspiration.

His response to a widely cited McKinsey claim, that around 95% of enterprise AI initiatives fail to deliver ROI, reflects the same logic. Pattisall doesn’t reject the spirit of the finding, but he challenges its interpretation.

“When you experiment, some things work and some things don’t. That’s not failure. That’s learning how to succeed.”

Seen that way, what looks like wasted investment often creates institutional muscle memory – knowledge that compounds over time, even if it doesn’t immediately satisfy CFO definitions of return.

The same lens shapes his thinking on commerce media and the creator economy. Pattisall sees confusion in the way the industry has sliced media into ever-narrower categories, when, in practice, it all serves different points along the same demand continuum.

“It’s all media,” he says. “Some converts directly. Some generates demand.”

Creators, meanwhile, reveal a different kind of fragility. Many, Pattisall notes, do not see themselves in the role for the long term.

“They all say, ‘I won’t be doing this in 10 years,’” he says. “It’s a launching point, not a career.”

That lack of permanence, he suggests, creates a gap in accumulated craft – the kind that only comes from decades of iteration and failure.

Which brings him back, unexpectedly, to CES itself. Amid the noise around agents, automation and platforms, Pattisall suspects the next phase will demand simplification as much as innovation.

“It would be very beneficial for the industry to move towards simplification, not just in structures but in how complex the marketplace has become.”

Perhaps, when the dust settles, future CES shows will spend as much time talking about craft and experience as they do about AI systems. Not a return to the past but a recognition that software, on its own, is never the whole story.