For thirty years the agency retainer encoded a simple assumption: marketing output is bottlenecked by skilled human hours, and you pay for those hours upfront so they're available when you need them. Every layer of agency pricing — billable rate cards, FTE allocations, scope-of-work documents, retainer minimums — is downstream of that assumption.
That assumption is no longer true.
AI agents execute continuously. They don't wait for the brief, the kick-off, the stakeholder review, or the next quarterly planning cycle. They produce, test, and optimise at machine speed — which means the bottleneck moves from execution to direction. Strategic clarity becomes the scarce input. Hours stop being the product.
This is not a productivity-tool argument. Productivity tools made existing humans 20% faster and the agency model survived. Agentic systems are categorically different: they remove the human from the execution loop entirely for a growing class of tasks. The work isn't faster — there is no human doing it.
What the retainer was actually paying for
If you decompose what a $40,000-per-month retainer covers at a traditional agency, you find roughly three things stacked on top of each other:
- Standby capacity. Senior strategists and creative directors who are available to your account, whether your account needs them this week or not. Their cost is amortised across your retainer regardless of usage.
- Production hours. The actual delivery of campaigns, creative variants, media buys, and reports. Most retainer hours are billed here. Most are also the cheapest, most fungible work the agency does.
- Coordination overhead. Account managers, project managers, status calls, weekly syncs, change-request workflows. Pure tax on the human-coordination problem of running an account.
An agentic system collapses two of those three. Production hours become near-zero marginal cost: an agent generating a hundred ad variants, refreshing audiences, and rebalancing bids across a portfolio doesn't bill differently from one doing ten of those things. Coordination overhead drops because the agent doesn't need a status call to know what it did last week — its working memory is its own logs.
What's left is direction. And direction does not scale on a billable-hour grid.
The retainer model fights the new shape of the work
Once production becomes cheap and continuous, the most valuable hours in marketing are the strategic conversations — the ones that decide what the agents should be optimising toward in the first place. Those conversations are bursty. They concentrate around launches, repositionings, market shifts, and crises. They don't fit neatly into a steady monthly retainer.
Worse, the retainer creates a perverse incentive. If clients are paying for hours and most hours are now zero marginal cost, the agency has every reason to keep finding ways to use up the hours — make-work that justifies the bill. We've watched this pattern play out across the industry over the last two years. Decks balloon. Reporting frequency increases. Workshops multiply. None of it improves outcomes; it improves utilisation, which is the only thing the retainer rewards.
"We were paying our previous agency for thirty FTEs of capacity. The actual work being done could have been completed by three people with the right tools. The other twenty-seven were a tax on a pricing model that no longer matched the technology."
That quote, from a CMO at a multi-property hospitality group in the UAE, is the version of the same conversation we have on every initial call.
Outcome-based pricing isn't new — but now it works
Performance-based pricing has been attempted for two decades. It almost always failed for the same reason: when execution is human, output is variable, and the agency cannot afford to take revenue risk on every account. The math only works when execution is reliable enough to amortise across a portfolio of clients.
Agentic execution is reliable. Not because the AI is perfect — it isn't — but because it doesn't get sick, leave, get poached by Meta, or hit a creative block at quarter-end. The variance in execution quality drops dramatically. And once variance drops, the agency can credibly take outcome risk.
That's why VIMDRIVE quotes engagements differently. We price against a small, fixed strategic envelope — typically a senior strategist's time, plus the agent infrastructure — and put the outcome at the centre of the conversation. ROAS, blended CAC, qualified pipeline, attributable revenue. The numbers we report are the numbers we're being measured on.
What this means for in-house teams
The retainer is a buyer-side artefact, not just a seller-side one. CMOs evaluate agencies through a procurement lens that asks "how many hours do we get?" because that's how the budget approval form is structured. We've seen perfectly capable in-house teams attempt to buy AI-native services and trip on their own procurement process — there's no line item for "agent capacity," only for "FTE-equivalents."
This will resolve over the next two to three years, but in the meantime there's a real asymmetry. Brands that work out how to procure outcomes rather than hours will move faster than their competitors. The first 18 months of agentic marketing in the Gulf are essentially an arbitrage window for that decision.
What replaces it
The model that replaces the retainer is closer to a managed-service contract than to a traditional agency engagement. You pay for:
- Strategic direction from a small senior team, billed against a fixed envelope — not hours.
- Agent capacity against your brand: the compute, the trained agent state, the integration with your data and channels. This is closer to infrastructure than to labour.
- Outcome variance: a performance component that pays the agency more when the agreed metric improves and less when it doesn't.
The line items shift, but more importantly the conversation shifts. You stop discussing how many hours of design were burned this month and start discussing whether the metric you actually care about moved.
That conversation is harder. Most CMOs are not used to running it. The agencies that learn to facilitate it — and to actually be accountable for the answer — will displace the ones still selling hours.
The honest objection
The honest objection to this argument is that not every brand is ready for outcome-based pricing. Some lack the data infrastructure to even measure the outcome cleanly. Some operate in markets — pure brand, long sales cycle, B2B with multi-quarter pipeline — where attribution to a specific quarter's spend is genuinely hard.
Fair. We don't claim the retainer dies tomorrow for every category. We claim it dies for performance marketing first, brand-with-direct-response second, and pure brand last. The order matters. If your spend mix is mostly performance, you should already be questioning the pricing structure of your current agency. If it's mostly brand, the conversation is closer to a five-year arc.
Either way, the assumption underneath the retainer — that human hours are the binding constraint — is the wrong assumption to organise the next decade of marketing around.