Agents everywhere, nobody coordinating them

In August 2025, Gartner predicted that 40% of enterprise applications would integrate task-specific AI agents by the end of 2026, up from less than 5% the year before. In media, we are already there. Google Ads has AI Max and its automated recommendations, Meta has Advantage+, TikTok has Smart+ and Symphony, Amazon has Performance+. Each one optimises inside its own platform with its own guardrails, and those guardrails do not look alike.

On Google, a Performance Max campaign accepts up to 10,000 negative keywords, brand exclusions and 50 search themes per asset group. The Search Partners opt-out is still in alpha, spotted in July 2026 on a handful of accounts, and household income exclusions have been seen on European accounts without being confirmed. On Meta, it is value rules, placement exclusions and frequency caps. On TikTok, a draft mode and a preview of creative combinations before launch, added in January 2026.

A small business with four accounts therefore ends up with four sets of rules, four vocabularies and four places where an exclusion can be forgotten. The list of competitor brands to exclude has to be entered four times, so does the spending cap, so does the brand safety rule. And when a platform adds an agent, it optimises in the platform's interest, with whatever rules it finds in the account.

At mediagentic, your accounts, your access and your rules are brought together in one place. I write your rules once and apply them on every platform. When one platform cannot apply them, I tell you right away, so we can decide together how to compensate. The platforms' agents stay in place, because they are good at what they do: they simply work inside a framework you wrote.