How AI Is Breaking The AOR Model
But AI has broken that dependency model.
Today, as generative tools, programmatic platforms, and machine-learning optimizers land inside the brand’s own walls, the AOR is under pressure to justify its existence. Scope is no longer a given. It’s a negotiation — and one increasingly dictated by proximity to data, speed of execution, and fluency By Megan C. Starkey
For decades, the agency-of-record (AOR) arrangement was a signal of stability. A Fortune 100 brand would entrust a single holding company with creative, media, and digital strategy — a bundled promise of reach, relevance, and continuity. It worked, mostly, because the tools of marketing were expensive and siloed. The agency was the access point.
in AI-driven infrastructure.
“We’ve gone from retainers to role clarity,” said one enterprise CMO. “If an agency doesn’t offer something we can’t replicate or outperform in-house, we rethink the contract.”
That rethink is happening across the Fortune 1000.
The Disintegration of Scope
AI didn’t just automate tasks — it exposed how many of them were redundant. Content versioning, performance creative, media buys, and even basic strategy decks are now executable by internal teams equipped with the right tools and a fractional data scientist.
It’s not that agencies aren’t valuable. It’s that value now needs to be visible.
And increasingly, the highest-value functions — personalization engines, first-party data models, cross-channel orchestration — require integration with systems most agencies don’t control.
So the power dynamic shifts.
In Bain & Company’s 2024 media study, 64% of brands reported restructuring their AOR relationships due to internal AI investment — either narrowing the agency’s scope or reassigning core functions in-house.
Three Emerging Models
What’s replacing the legacy AOR is less a monolith than a mosaic. The most forward-looking brands are gravitating toward one of three configurations:
The Embedded AI Partner A few elite agencies have gone all-in on AI — not just offering tools, but integrating directly into a brand’s stack. Publicis, for instance, built CoreAI as a platform layer to ingest brand data and co-produce content, optimization, and media recommendations. It’s a new kind of embedded vendor — not an executor, but a systems layer.
The Hybrid Execution Model Here, brands insource the strategic core — AI-driven segmentation, insights, and even prompt-based creative — and use agencies for localization, experiential, or niche activations. The agency becomes an extender of brand logic, not its architect.
The Displaced Vendor The most vulnerable agencies are those offering services that generative AI has commodified: static creative, generic insights, templated campaigns. These firms may retain contracts — for now — but they’re in quiet scope collapse.
Not every brand will want to internalize this complexity. Some are actively seeking partners who simplify the stack, not just automate it. For those agencies, the opportunity is to become AI translators — not just executors.
What This Means for Leadership
This is not just a procurement shift. It’s a strategic one. What’s at stake isn’t just efficiency — it’s control over how a brand thinks, moves, and speaks in the market.
For CMOs, the question is no longer “Do we trust the agency?” but “What do we need to own in order to compete?”
And not all brands are ready to move upstream. Internal AI capability takes investment, structure, and executive clarity. The question isn’t just “Can we use these tools?” — it’s “Are we architected to extract value from them?”
For agencies, the path forward is simple, but difficult: stop selling execution. Start selling systems fluency. Productize your AI. Prove your value not through polish, but through performance.
“In the next year, we’re going to see more brands issuing RFPs that include questions like: What’s your generative pipeline? How do you integrate with our CDP? Can your insights layer talk to our media stack?”
If an agency can’t answer — or even understand — those questions, they won’t be an AOR much longer.
AI isn’t just changing what agencies do. It’s changing what agencies are.
This article is part of CMO/AI, a newsletter for marketing leaders navigating the structural shift AI is bringing to growth and revenue functions. Based on the forthcoming book, AI Unlocks: A Playbook for Marketing Leaders Navigating AI Adoption (2026).
Published by Megan C. Starkey, founder of RBD., an advisory firm that helps enterprises turn AI into a scalable, revenue-driving capability.
