As an agency alum who has worked across both the buy side and sell side – and now at PubMatic – I’ve spent years in rooms with agency buyers, from independent to major holding companies. These are people responsible for moving millions, sometimes hundreds of millions, in media spend. They’re smart, experienced, and skeptical in exactly the right ways.
That skepticism shows up quickly when agentic AI enters the conversation.
How do I know it won’t just spend all my client’s money in the wrong places? How do I explain the decisions it’s making? How do I stay in control when I’m ultimately the one responsible for the outcomes?
Those are absolutely the right questions and the same ones I’d be asking if I were in their position.
What we actually built
The agents we’re deploying through AgenticOS don’t operate autonomously in the scary sense of the word. They operate within guardrails that you set. They surface recommended actions for your approval, and then generate detailed reporting on every impression, every placement. What you’re losing is the friction, the manual labor — not the control.
Think about the pre-call prep for a client meeting. Right now, that probably means logging into a backend UI, pulling reports, translating data into something a client can follow, and hoping nothing has shifted since the last time it was pulled. That’s time that could be spent on strategy, on preparing to wow the client with an outside-of-the-box solution, or on getting them amped for the next campaign. Right now, that’s lost on manual overhead.
With an agentic workflow, that same prep looks like opening your own AI interface — one that talks back in plain language — and asking: How’s my campaign pacing for Client A? Where has the majority of spend landed so far? Where else can we move those dollars?
And then walking your client through the same interface in the meeting. No asking them to interpret algorithms or spreadsheets. Just a clear, shared view of what’s working and what to do next.
That’s the kind of transparency most buyers have never actually had.
We’ve actually been here before
When PubMatic launched Activate nearly three years ago, the learning curve was real.
We asked buyers to do something genuinely different. Instead of transacting on the buy side, we asked them to transact on the sell side – like bypassing your real estate agent to hand the money directly to the homeowner. More efficient, more direct, but it’s a big purchase, and it’s not how anyone learned to do it.
The benefits outweighed the hesitation – real cost efficiency and real verified performance. Today, every major holding company is buying strategically through Activate, and the platform saw 100% growth in activity last year.
Now we’re asking for a different kind of trust.
Let the agent handle the negotiation, the paperwork, the inspections and approvals. It’s a bigger ask. We know that. These are your client relationships on the line, not just your workflow.
But here’s how we built AgenticOS: we started from the assumption that you should never have to take our word for it. Every action is auditable. Every recommendation requires your approval before it is executed. Your guardrails are the guardrails — the agents operate within the boundaries you set, not ours. And all of it surfaces in an interface you already work in, in plain language, not a new dashboard you have to learn.
We’re meeting agencies where they are and building this out in the open, because, ultimately, transparency is the foundation of trust.
Why the next 12 months matter
The buyers who engage with agentic AI now — seriously, strategically, with the right questions — are going to help shape how this ecosystem develops. What standards look like. How governance works. Where the open internet fits in.
The buyers who wait will inherit decisions made by others.
We wrote this for you
We built The Agentic AI Blueprint because our partners deserve more than buzzwords. It’s a grounded look at what agentic advertising actually means in practice. We explore how it works, how multi-layer optimization compounds over time, how governance and transparency have to be foundational rather than an afterthought, and a practical framework for evaluating who you should be building with.
We’re intentionally building this out in the open because the buyers closest to the problem should help shape the solution.
Download The Agentic AI Blueprint and reach out to me to get started on testing and building together.