Agentic advertising changes where marketing decisions get made, how work gets executed, and what clients will actually pay agencies for. McKinsey’s recent work on agentic AI is direct about it: these aren’t better chatbots. They’re systems that act across entire workflows with defined goals, tools, and permissions. McKinsey
That matters because the traditional agency model still makes too much of its money from managed labor, production volume, and process overhead. Agentic systems will compress all three but that’s only a problem if your agency is still selling output when clients want systems, governance, and judgment.
The work is moving closer to the budget owner
The first real change is where the work happens. Media planning, segmentation, creative versioning, optimization, and reporting are moving closer to the brand’s own data, tools, and budget controls. McKinsey’s marketing research points to a future where AI-enabled workflows turn marketing into a continuous operating loop rather than a sequence of isolated campaigns. McKinsey
That reduces the need for agencies to sit in the middle of every task. It also raises the value of people who can design the system, set the rules, and know when automation is helping versus when it’s creating risk. The work doesn’t disappear, it changes shape.
For agencies, the labor-heavy parts of the model come under pressure first. Clients won’t keep paying premium fees for repetitive manual processes when systems can do the work faster, cheaper, and at scale. The real opportunity is to move up the stack, not protect the parts of the model that were already aging out.
This is good for clients, and eventually good for agencies
Clients have wanted this for years, even if they didn’t describe it this way. They want faster iteration, better personalization, cleaner attribution, and less drag between strategy and execution. McKinsey argues that the real advantage now goes to organizations that can learn faster, personalize at scale, and optimize continuously. McKinsey
That should force agencies to become more valuable and push them away from selling managed labor and toward selling judgment, design, governance, brand stewardship, and accountability. The disruption is healthy because it exposes the parts of the agency model that were already weakening.
A lot of agencies know this as they’ve been overstaffing production, underpricing strategy, and hiding weak economics inside activity. Agentic advertising makes the line much clearer between work that can be systematized and work that still depends on human judgment.
The real risk is not job loss, it’s category confusion.
The laziest version of this conversation is that AI is taking agency jobs. Some jobs will shrink and some roles will be redesigned. The bigger risk is that agencies misread the category shift and keep presenting themselves the old way.
If an agency still leads with deliverables, team size, responsiveness, and process, it’ll look more replaceable every quarter. If it leads with commercial judgment, creative quality, governance, decision design, and accountability, it can become more valuable in a market where clients are under pressure to automate safely.
That distinction is already showing up in the data as Salesforce’s latest State of Marketing found that 75% of marketers are already using AI, 81% would trust AI to help interact with customers, and yet 84% still admit to running generic campaigns. The problem is weak orchestration, weak data, and weak judgment. Tools aren’t the gap. Salesforce
Agencies need to move away from “we use AI too,” because that’s table stakes and toward “we know how to make this work without damaging your brand, economics, or customer experience.”
Publishers are in this story too
Publishers shouldn’t assume this disruption only lands on agencies as it affects them directly.
If more planning, targeting, optimization, and reporting move into AI-mediated workflows, publishers need to become easier for those systems to trust, interpret, and buy against. That means stronger metadata, better audience signals, cleaner inventory, and clearer governance around how AI is used across content and advertising environments.
Trusted context may become more valuable as synthetic content spreads and premium publishers can position themselves as safer ground for brands. The same market that rewards speed will also reward trustworthy environments, especially when automation starts making more buying decisions upstream.
Governance is no longer optional
This is the part the industry still likes to treat as an afterthought but in my opinion is that it shouldn’t be this way.
Agentic systems don’t just generate content, they make decisions, call tools, trigger actions, and create consequences at speed and once that happens, accuracy is only one question. The harder one here is who’s accountable when the system acts, and what rules existed before it did?
In marketing and advertising, that means governance has to be part of the operating model. Brands and agencies need clear rules around what data an agent can access, what claims it can make, what spending authority it has, how outputs are reviewed, and what escalation triggers exist. Without that, speed quickly turns into liability.
The industry is beginning to respond and the IAB’s AI Transparency and Disclosure Framework is a useful step. It gives advertisers, agencies, publishers, and platforms clearer standards around when disclosure is needed and how to handle responsible AI use in advertising. IAB
Pricing has to change too
The old agency pricing model was already under strain before agentic AI arrived. Too much pricing still depends on hours, headcount, and production volume and agentic systems challenge all three.
Agencies should be moving pricing toward responsibility, outcomes, and governed systems and that doesn’t mean every engagement becomes pure value pricing. It means the fee should increasingly reflect what the agency is accountable for: system design, strategic judgment, performance oversight, integration, brand safety, and governance.
This is one reason I see the change as healthy because it forces agencies to stop hiding weak economics inside effort. If a machine can do repetitive work, the agency has to explain what remains and why it matters and that should improve the business, not weaken it.
What agencies should do now
I think that what people need to do first is separate strategy from production and treat them as the different kinds of value they are.
Second, identify which workflows are actually differentiating and which are just expensive habits because a lot of internal agency processes survive because nobody has challenged them, yet.
Third, build governance into the service itself, not as an add-on. Clients will need help with audit trails, approval logic, disclosure, safety controls, and escalation paths and iyou’re helping a client build or run an agentic marketing system, that governance layer is part of the value.
Fourth, redesign pricing around accountability which means if the agency is helping a client build and govern an agentic marketing system, the fee should reflect that responsibility.
Fifth, train teams to become better judges. The machine can handle more of the doing but the human advantage shifts toward deciding what’s worth doing, what shouldn’t be automated, and where the brand is exposed.
Parts of the agency model will die. The right parts.
Agentic advertising will disrupt the traditional agency model and it should. The old model that has been used for decades rewarded labor-heavy execution, process management, and volume.
The next one will reward system design, trustworthy data, strong judgment, clear governance, and the ability to connect automation to real business outcomes without creating reputational damage and I don’t see that as a tragedy but an overdue move.
The agencies and publishers that adapt early will become the ones clients build around.
Travis Pomposello is a seasoned media and creative executive who spent nearly four decades inside the machinery of high performance media, from iconic TV networks to scaling multimillion dollar creative agencies. Today he is a New York based global mentor for creative agencies.
