Forrester and the 4As (American Association of Advertising Agencies) say US marketing agencies now rank AI vendors by ecosystem integration, not raw model performance, and Google was already built for that question.
The leaderboard shuffle in agency AI preferences looks, at first glance, like a horse race: Google has overtaken OpenAI, Adobe, Anthropic, and Microsoft as US marketing agencies' preferred AI partner for the first time since 2024, according to new joint research from Forrester and the 4As. The more durable story is the question behind the ranking. Agencies are no longer grading vendors on which model is smartest. They are grading them on which ecosystem best connects data, creative, activation, and measurement into a single working system, and Google was already built for that question.
Forrester frames the new criterion as "marketing orchestration at scale": an end-to-end marketing operating system rather than a stack of point tools. The shift is procurement logic, not a referendum on model benchmarks. Agencies that buy AI today have to make it plug into the work they already do, which is media planning, audience targeting, creative production, and post-campaign measurement. A vendor that owns adjacent layers of that workflow, from Analytics 360 to Meridian to Asset Studio to Gemini, Imagen, and Veo, and from Search and YouTube to Display, can answer that question with an integration diagram. A vendor that ships a frontier model and a clean API has to convince a procurement team the rest will follow.
The 2026 Cannes Lions International Festival of Creativity, the annual global gathering of the ad industry on the French Riviera, gave the moment its timing. The festival's opening week coincided with the publication of the ranking, and the floor talk this year has been less about which model is winning benchmarks and more about which AI ecosystem is winning the operating stack. Google is leaning into the occasion heavily. OpenAI, Anthropic, Microsoft, and Adobe are also activating on the Croisette. Preference, on the record, has moved.
There is a real critical read the ranking does not erase. Consolidation of this kind reduces vendor choice for agencies and their clients, raises switching costs once an operating system is in place, and gives the incumbent an integration moat that newer model challengers cannot easily replicate even if their raw model quality is competitive. Forrester's "ecosystem advantage" framing is the firm's own interpretation of the data, not a Google talking point, and the lock-in risk is the other half of the same story.
The watch item is whether the criterion holds. If enterprise AI buying in marketing stays orchestration-shaped, Google's lead narrows only when a competitor matches the integration depth, and the model-quality race stops mattering on its own. If procurement reverts to model scoring, today's leaderboard will reset on the next benchmark.