That is the sentence we keep coming back to. It is the sentence the holdcos already know. It is the sentence the platforms are quietly betting you won’t say out loud until it’s too late to do anything about it.
“Independents don’t need a platform’s agent. They need their own.”
The moment
Look at what shipped in the last twelve months.
Amazon Ads launched its agentic buyer inside the DSP. Google rolled agentic placements into Performance Max. The Trade Desk leaned harder into Koa and Kokai as decision layers, not just optimization knobs. Meta turned Advantage+ from a campaign type into a default. As of May 2026, every walled-garden and major DSP either has an agent in market or has one announced.
Every one of those agents is biased toward the inventory the company that built it sells. That isn’t a moral failing. It’s a business model. Amazon’s agent will spend more on Amazon. Google’s agent will spend more on YouTube and Google-owned placements. TTD’s agent will lean into TTD-curated supply paths. Meta’s agent will not, under any condition, send a dollar to TikTok.
At the same time, the holdcos are not standing still. Omnicom announced its agentic stack. Publicis is building inside Epsilon. WPP is making its third attempt at a unified intelligence layer. IPG, Dentsu, Havas, Stagwell — every one of them has a press release and a roadmap. By 2028, the majority of holdco clients will be bought by a holdco-owned agent.
Twelve months ago, none of this was real in production. Agents were demos at IAB. Today every major buyer in the market has at least one in market, in beta, or shipping next quarter. That is the speed of the shift. Anyone telling you it’s a multi-year decision is selling you on a multi-year decision.
So where does that leave the independent agency?
The trap
Two doors. Both are bad.
Door one:adopt the platforms’ agents. You become a frontend on someone else’s biased stack. Your client’s media plan is shaped by whichever platform’s agent gets the most weight in your workflow. You can’t honestly tell that client you’re independent anymore — you’re just the integrator.
Door two:do nothing. Keep buying the way you’ve bought for the last five years. You watch a holdco pitch your account with a five-minute demo of an agent that auto-builds a plan, auto-buys against it, and auto-reports on it. You lose the pitch on operational sophistication, not strategy.
Neither door ends well for an independent agency that wants to still be independent in 2028.
There is a third door. It is the one the holdcos and the platforms would prefer you don’t notice yet.
The third path
Independents don’t need a platform’s agent. They need their own.
An agent built around what makes your agency yours. Your channel strategies. Your definition of premium media. Your brand-safety rules. Your audience playbook. Your frequency and pacing logic. Your performance history per client.
Codified once. Executed on every impression. Owned by you.
This is not theoretical. The components are all here. Open exchanges are still open. SSPs still take direct connections. Curation deals are easier to set up than they’ve ever been. The work that used to take a 12-person trading desk now takes an agent that has read your playbook and is authorized to buy against it, with a human on the approval line for anything that matters.
The only thing missing has been the tool that lets a 30-person agency stand up that kind of agent without writing the software themselves. That’s the gap we built into.
What this looks like in practice
The shift is more boring than the headlines make it sound, and that is the point.
A senior planner sits down on Monday with a client brief. The agent has already ingested the agency’s playbook: the channels this client has historically bought, the publishers on the premium list, the categories on the block list, the frequency caps, the dayparting, the negative-keyword themes, the segments the agency has built up over years. The planner doesn’t re-explain any of that. The planner explains the client’s quarterly goal and the budget.
The agent proposes a media plan. The planner reviews, edits, approves. The agent buys against it across whichever exchanges and curation deals fit the premium definition, with rules that the agency wrote, not rules that the DSP shipped. The planner sees the next decision queue in the morning — bid changes, creative rotations, pacing adjustments — and approves what should ship.
Nothing the agent does is invisible. Nothing the agent does is locked in by a vendor whose margin depends on it. The agency keeps the relationship, keeps the data, and keeps the upside.
The ten-person trading desks the holdcos used to run are now eight-person trading desks. The thirty-person trading desks the holdcos used to run are now ten-person trading desks. The independent agency that wires this up correctly runs a similar plan with a planner-and-a-half and a smarter tool. That is the structural shift. That is the year independents catch up on operations and pull ahead on incentive alignment.
What Halliard is doing
Halliard is the buying agent built for independent agencies.
Not a DSP. Not a holdco product. Not a retrofit of a SaaS tool that used to do something else. An agent that you train on your agency’s playbook and point at the open market on your client’s behalf.
A few things that matter:
- Flat take rate.No share of media. No hidden curation cut. The agent is on your side because we’re paid by you, not by the supply path. An early agency partner has been running on this model since their pilot — they tell the story better than we can, and more of it will be published later this year.
- Open market. No house DSP. No preferred SSP. The agent shops the exchanges that match your premium-media definition for the client in front of it.
- Your playbook, not ours.We don’t tell you what “premium” means. You tell the agent. We give you the surface to codify it once and re-use it across clients.
- Human-in-the-loop. Every meaningful buy decision flows past a human. Pilots stand up in weeks, not quarters, and the kill switch is yours.
We are onboarding a small pilot cohort of independent agencies through the back half of 2026. The cohort is small on purpose. We want the early agencies to shape what the agent ingests and how it reports, because they’re the ones who will be running it for the next decade.
The choice
Every independent CEO in this country is making this decision right now, whether they realize it or not. The agent question is the next strategic-vendor question, on the scale of “which DSP did you bet on in 2015.” Get it right and your shop has a structural advantage going into every pitch for the next three years. Get it wrong, or wait too long, and your shop becomes a UI on someone else’s economics.
We think this is the moment independents lead. Not because they have to. Because they’re the only group of agencies in the market whose incentives are still aligned with the client’s. The holdcos have principal margin to defend. The platforms have inventory to sell. The independents have the client. That is the asset.
The agent should reflect that.