The Creative Agency of 2028 · How the Best Are Already Operating

I ran a creative agency for over a decade. The agencies that are going to triple their margins by 2028 are not the ones doubling down on headcount or chasing the next AI tool everyone is already buying. They are the ones rebuilding the operating model itself. And the ones that survive are doing it right now, quietly, while their competitors are still debating whether to add a prompt engineer to the team.

This essay is the operator's view of what is changing, why the agencies with ten-plus clients running multi-product portfolios feel the pain first, and what the working model looks like for the agencies that come out ahead. I closed my own agency to build this. That is how confident I am about where this is going.

What changed between 2018 and 2028

In 2018, a competent creative agency could service ten clients with a team of twenty and run a healthy margin. The math was clean: a few strategists, a creative team, an account layer, a production layer, and the workflow held. By 2024, that same agency was running thirty creative variants per campaign instead of three, briefing four sub-agencies for asset production, and renegotiating retainers every quarter because clients had started measuring agency output in volume instead of insight.

By 2026, the volume expectation tripled again. The market is asking creative agencies for the output of a 2018 agency at ten times the velocity, across four times the platforms, in three more languages. The agencies that responded by hiring more people watched their margins evaporate. The agencies that responded by buying more tools watched their teams burn out faster. The honest read on what most agencies look like today is in our piece on how agencies became content factories and what that has cost the strategy layer.

The agencies surviving 2026 with healthy margins did one of two things. Either they collapsed their scope and went niche, or they rebuilt the operating model underneath. The first path is defensible. The second path is the one that compounds.

Where the pain concentrates · agencies with ten-plus multi-product clients

The agencies feeling this hardest are not the small shops with three boutique clients. They are the mid-sized agencies running ten or more clients, each of them brands with multiple product lines, multiple buyer personas, and multiple regional markets. That is where the volume problem becomes a structural problem.

Consider the math. A typical client in this profile runs four product lines, three buyer personas per product, and two priority geographies. That is twenty-four creative territories, before you factor in seasonal campaigns, performance variants for paid social, or any kind of platform-specific adaptation. A campaign brief for one client in this profile is, in practical terms, twenty-four parallel campaign briefs.

Industry research from Gartner's 2025 marketing operations report shows that brands with more than three product lines now allocate 38% of their marketing operations budget to creative production coordination alone — a line item that did not exist as a category five years ago. Agencies are absorbing that coordination cost, often invisibly, by burning senior strategist time on what is fundamentally a workflow problem.

The hidden trade-off · strategy sacrificed to protect margin

Here is the part that does not show up in any agency report, but every agency owner I have spoken to recognises immediately. When the volume demand on a multi-product client starts to outrun the team's capacity, the first thing that gets cut is strategy depth.

Not officially. Officially, the strategy work is still on the SOW. But in practice, the senior strategist who used to spend three days per client per month doing positioning work gets pulled into production review, because production is on fire and clients are asking where the next round of assets is. The strategy layer thins out. The work that goes out the door is still on-brand, still on-time, but the thinking behind it gets shallower with every cycle.

The agency owner sees this. The agency owner usually caused it, because the alternative was bringing in another senior strategist at a cost the margin could not support. So the trade-off becomes: protect the margin slightly, at the cost of giving the client a thinner version of the strategy they are paying for.

The first time, the client does not notice. By the third quarter of this pattern, the client starts asking different questions in QBRs. By the second year, the client is in the market for a new agency. The math on this is brutal and we walked through it in detail in the CAC-LTV spread of agency retention — the cost of losing a multi-product client almost always exceeds the cost of the additional strategist the agency refused to hire.

The infinite stress loop · why this becomes a cultural problem

The trade-off described above is not a one-time event. It is a loop, and once a multi-client agency enters it, the loop self-reinforces. The senior team feels it first. The strategist who can no longer do real strategy starts looking for somewhere else to do real strategy. The account director who keeps absorbing scope creep without renegotiating retainers starts burning out. The owner, watching margin tighten, hesitates to invest in anything that does not show immediate billable output.

By the time it reaches the team layer, the loop has compounded into something close to permanent operational stress. Harvard Business Review's research on creative burnout in agency environments tracks attrition rates in mid-sized creative shops at over 32% annually for production roles, with senior strategist tenure dropping below two years industry-wide. The agency is paying an invisible tax in turnover that always ends up larger than the savings it tried to protect on the margin.

And underneath all of this, the agency owner watches the same pattern play out across every client account. The same scope creep. The same strategy thinning. The same eventual churn-to-competitor. The honest experience of running a multi-client agency in 2026 is the experience of trying to plug a leaking boat with one hand while rowing with the other.

What the agency of 2028 actually looks like

The agencies that come out of this period with tripled margins do something specific. They stop trying to fix the workflow problem with more people, and they stop trying to fix it with more tools. Instead, they fix it by replacing the operating model underneath the agency itself.

Concretely, the agency of 2028 operates as a three-layer system:

• The brand intelligence layer. A persistent, structured representation of every client — voice, product context, audience segments, prior campaign learnings, guardrails. The senior team builds this once per client and keeps it current. The system reads from it on every subsequent task. We walked through the architectural shape of this layer in detail in our piece on the Creative Graph.
• The production layer. The execution of variants, adaptations, asset generation, and platform-specific compliance happens inside the operating system, not inside a Slack channel of fifteen people. The production layer reads from the brand intelligence layer, produces output that is on-brand by construction, and ships to the delivery layer without losing context.
• The strategic layer. The senior team — the strategists, the account leads, the creative directors — spend their time on what they were hired to do. Positioning, narrative, deep audience work, P&L conversations with the client. Not production triage.

The agency of 2028 is not smaller. It is differently shaped. Roughly four senior operators producing what a 2018 team of twenty used to produce, with strategy depth intact and a healthier margin, because the production cost has been collapsed into the system layer.

The math of tripled margins · honest unit economics

Let me put numbers on this, because the abstract version is easy to wave away. Take a typical mid-sized agency: ten clients, $4M annual revenue, twenty-person team, twelve percent net margin. The 2018-era operating model.

The 2026 column is the agency you already know — same team, more pressure, eroded margin, strategy depth collapsed. The 2028 column is what happens when you accept that the bottleneck is not the team's talent — it is the operating model. Revenue is intentionally slightly lower in 2028, because the agency is servicing fewer clients with more depth. Margin almost triples. Strategy depth quadruples.

The agencies that get to that 2028 column are the ones that stop treating operating-system infrastructure as an overhead line and start treating it as the load-bearing wall of the business. We argued the broader version of this case in AI marketing tools vs AI marketing systems.

What gets retained · the client side of the equation

The agencies running this model do not just protect their own margin. They visibly improve the work they deliver to the client. And the retention math changes.

A client who feels their agency is doing strategy work — real strategy, not re-formatted brief work — renews. A client whose agency keeps the brand intelligence layer fresh, surfaces insights from past campaigns, and shows up with provocations rather than asset previews, does not shop the account out at the end of the contract year. Deloitte's 2025 agency-client relationship study shows that agencies with documented strategic-depth investment lose clients at less than half the rate of agencies operating in pure production mode.

Multi-product clients in particular are sensitive to this. They notice within two quarters whether their agency is operating a true intelligence layer across their products, or whether each product brief gets handled by a different account team with no shared memory. The agencies that win the long contracts in 2028 are the ones where the answer is unambiguously the former.

The operating model in detail · what to build first

For agency owners reading this who want a concrete starting point, the sequence that works is the one we have seen play out at over a dozen agencies in the last twenty-four months. It is ordered deliberately; the order matters more than the speed.

• Step one · Build the brand intelligence layer for two clients. Not all ten. Two. Pick the clients with the most product complexity. Get the intelligence layer captured well enough that someone who has never worked on that account can produce on-brand first drafts from it.
• Step two · Move production for those two clients into the system. Asset generation, variant production, platform adaptation — all happening inside the operating system, not in a parallel Slack-and-Figma flow.
• Step three · Redeploy the freed-up senior time toward strategy. The two clients should be getting four-plus days per month of senior strategy work within ninety days. The QBR conversations should noticeably shift — from production status to strategy provocation.
• Step four · Renew those two clients at a higher rate. The retention conversation is different when you can show the QBR record of the prior year. Use the renewal margin to fund the next two clients onto the system.
• Step five · Wind down the production-only roles deliberately. Not as a layoff event. As a structural shift, communicated honestly, with timelines that respect the people doing the work. Some of those roles convert to senior strategy with retraining. Some do not.

Steps one through three take roughly six months. Step four lands in the next renewal cycle. Step five is the eighteen-month horizon. By month twenty-four, the agency looks different from the inside and noticeably better from the outside. Margin recovers. Strategy depth is restored. The client churn rate flattens, then improves.

Why I closed my agency to build this

I founded Wizerlink as a creative agency over a decade ago. We hit the volume problem described in this essay around 2023. I tried the obvious things first — more production hires, more AI tools, more sub-agency partnerships. Margin kept tightening. Strategy depth kept thinning. I watched senior people leave because the work they had been hired to do was no longer the work they were spending their days on.

We rebuilt the operating model underneath, the way I have described here. The numbers we published in our case write-up — 40% lower operational cost, 16x faster delivery on the same campaign scope — are real, and they came from running this model for two years inside the agency. The agencies I have spoken to since have all asked the same question: how do I do this without rebuilding the operating system layer from scratch?

I closed Wizerlink to answer that question by building Hi Luca. The agency of 2028 is real. The operating model exists. The question for an agency owner reading this in 2026 is not whether to make the shift — it is whether to make it now, while clients still have patience for an agency that is mid-transition, or to make it in 2027 when the multi-product clients have already moved to agencies that finished the shift first.

We built Hi Luca specifically for this operating model. If you want a practical sense of the seat in the system reserved for the agencies making this move first, we documented it in our Founding 20 program — twenty agency partners we are working with this year on exactly the transition described above. The broader operator perspective on what life looks like inside an agency once the operating model is in place is documented in Before / After Hi Luca.

McKinsey's 2025 State of AI in Marketing report projects that by 2028, the agencies that have completed the operating-model shift will represent less than fifteen percent of mid-sized agencies in the market — and that those fifteen percent will capture roughly sixty percent of the multi-product client portfolio. The window for being in that fifteen percent is open right now. It will not stay open indefinitely.