The Title That Took a Decade to Arrive
A decade ago, “AI” in a media plan was a modest bidding layer that nonetheless worked. In 2026, the work has matured into an institutional mandate, visible in the emergence of the Chief AI Officer across major holding companies.
I first wrote artificial intelligence into a media plan in 2016, in a Manila office of Omnicom, and the phrase landed on the page with the gentle absurdity of all early language. The campaign was for SC Johnson Baygon. The strategic problem was ordinary — drive in-store participation in a flood-season promotion, the kind of brief that in 2016 would have been answered with mass-reach digital and a hopeful coupon. We did something else.
What we built, in the vocabulary the industry now uses, was a real-time event-driven targeting pipeline that fused three data layers the market had not yet learned to fuse. The first was Project NOAH, the Philippine government’s open-source flood and disaster monitoring system, which published live hydrological and weather data through public APIs. The second was Factual, a third-party location intelligence partner whose dataset gave us point-of-interest coverage of every Robinsons supermarket carrying the promotional product. The third was the demand-side platform layer, where we could express bid logic against geofenced polygons in something close to real time. We composed these three sources into a single decision rule: when flood conditions intensified in a barangay whose population sat within a defined catchment radius of a participating Robinsons store, the campaign pushed promotional creative to mobile users inside that polygon, with a call-to-action calibrated to immediate household need.
The campaign delivered above benchmark on reach and click-through. That is not the part of the story that interests me ten years later. The part that interests me is what we were actually doing, which had no name in 2016. We were running a multi-source contextual AI pipeline on a single shopper-marketing brief. The pipeline ingested public government data, commercial POI data, and proprietary audience signals; it applied a decision logic against fused inputs in something close to real time; and it triggered creative deployment at a granularity the platforms of the period were not yet built to expose. The architecture was hand-built. The categorical name for it arrived later.
That same year, on a McDonald’s Father’s Day brief, the team took a single sixty-second television commercial and built around it what would now be called a dynamic creative optimisation deployment. We segmented the male thirty-to-fifty-five audience into three psychographic cohorts — first-time fathers, fathers of millennials, fathers of generation-X — and produced three creative cut-downs from the master asset, each targeted against a custom matrix of keywords, life-stage signals, geographic clusters, and day-parting windows. The point was not the creative variety. The point was that audience-level creative personalisation, in 2016, required someone to manually design the taxonomy, build the variants, write the targeting logic, and trace the performance back to the cohort. We did it because the platforms had not yet absorbed the work into a one-click feature. Today they have. The capability is unremarkable. In 2016 it was not.
I describe both campaigns precisely because the trade press of the period rarely did. The work I and others were doing in 2016 and 2017 and 2018 was operationally serious and organisationally invisible. We were building algorithmic capability inside country offices, inside trading desks, inside performance practices, in pockets where the work fit a profit-and-loss owner’s quarterly need. We had no charter to make the work travel. We had no title that signalled what we were actually doing. The plan said programmatic. The work was structural.
Looking back, that period marked the start of a decade-long process by which artificial intelligence migrated from the margins of media operations to the centre of how the work is conceived. The migration was uneven. It moved faster in some markets than in others, faster in some disciplines than in others, and faster inside some agency networks than the trade press suggested. But it moved continuously, and the people who carried it were largely uncredited, because the work happened inside rooms whose existence the deck above them did not advertise.
The most interesting feature of 2026 is that those rooms now have a name on the door. The Chief AI Officer has arrived inside the holding companies as a peer of the Chief Strategy Officer and the Chief Product Officer. The title was experimental two years ago and is structural now. The pattern is visible across each of the major networks — group-level AI leadership has been appointed across WPP, Publicis Groupe, Omnicom, and Interpublic, in most cases reporting into the office of the chief executive.[1] The pattern is too consistent to be coincidence, and it is the most important governance development in the agency business since the formation of the trading desks fifteen years ago.
What the Chief AI Officer class does, that the prior decade of practitioners could not, is carry the conversation upward. The Baygon pipeline and the Father’s Day deployment were both deeply operational and structurally invisible — they sat inside campaign reports and end-of-quarter wrap decks, not inside the corporate strategy documents that would have travelled upward through the network. The Chief AI Officer has charter to make that travel happen. The role exists to make the work travel — across markets, across disciplines, across clients, and most importantly upward, into the conversations where capital allocation and client relationships are decided. That is not a small thing. That is the structural piece that was missing for a decade.
I notice, watching this cohort form, two patterns worth naming. The first is that the strongest Chief AI Officer appointments are coming from inside the buildings rather than from outside them. The networks have learned, expensively, that hiring an AI executive from a pure-play technology company produces a translation problem of the kind that consumes the first eighteen months of the appointment. The internal candidates — the ones who came up through the trading desks, the analytics practices, the performance teams, the regional innovation functions, the people who in 2016 were fusing government APIs with DSP polygons because the brief demanded it — arrive with the institutional fluency that lets the work begin on day one. The networks are choosing wisely.
The second is that the role is forming alongside — not in competition with — the independent practitioner ecosystem that has emerged outside the networks over the same period. I run a small AI advisory and operating practice across four jurisdictions, the Dubai entity being Infinite Momentum Holdings and the Singapore entity being Acceler8 AI Pte. Ltd., under the AI-Driven Media ROI™ proposition. Two years of operating that practice has produced a counterintuitive observation. The independent practices and the holding-company Chief AI Officers are not adversaries. They are complements. The independent practices carry experimentation velocity and cross-client pattern recognition that the in-house function cannot generate at the same speed. The in-house function carries client trust, contract gravity, and the ability to deploy capability at the scale of a multi-hundred-million-dollar account portfolio. The strongest commercial AI work of the next five years will happen at the seams between these two structures, not inside either of them alone.
What I find moving about this moment is that the cohort now carrying the work is the right cohort. They are operationally honest about what the technology can and cannot yet do, and they are politically careful in a way that suggests they understand the difference between what an organisation can announce and what it can absorb. They have grown into a role the industry needed someone to grow into. The title took a decade to arrive. The work it describes was always already happening. The next decade will be more serious because the title is finally on the door.
[1] WPP, Publicis Groupe, Omnicom, and Interpublic have each made group-level AI leadership appointments between 2024 and 2026, with varying titles (Chief AI Officer, Global AI Lead, Group AI Officer) but consistent reporting lines to the CEO office.
By Jona Oboza — Founder & CEO, Acceler8 AI.