Chapter 5: The New Agency Economics
AI creates one very awkward commercial problem.
The better you get, the faster you get.
The faster you get, the harder it becomes to justify selling time.
Which means if your agency gets efficient but keeps billing by the hour, congratulations: you’ve built a margin shredder.
This is where a lot of agencies are heading.
Operationally faster.
Commercially unchanged.
Quietly less profitable.
That doesn’t hold for long.
The old model loved friction
The traditional agency model worked because production was slow, specialist and labour-heavy.
More people created more output.
More hours created more revenue.
Longer timelines made the work feel more serious.
AI messes with that beautifully.
Research gets synthesised in hours. First drafts appear in minutes. Visual territories expand at speed. Rollout becomes less painful. Production effort compresses.
The value shifts away from:
“How long did this take?”
Toward:
“How good is the thinking?”
That is a massive change.
And most agencies have not priced for it properly yet.
Clients were never really buying hours
Clients bought hours because hours were the easiest thing to measure.
What they actually wanted was:
confidence
momentum
judgement
creativity
strategic clarity
quality
reduced risk
Hours were the proxy.
AI exposes how weak that proxy always was.
A sharp agency can now create stronger outcomes in less time.
That should increase value.
Not reduce it.
But only if the agency knows how to sell the value.
The value moves upstream
When AI makes execution easier, the premium moves upstream.
Clients still need people who can:
define the problem
spot the real opportunity
choose the right idea
know what to remove
protect the brand
manage risk
make work distinctive
steer the room
That is where the money should sit.
Not in the hours spent resizing the same asset into 19 formats, with everyone pretending this is a noble use of human life.
AI compresses execution.
So agencies need to charge for judgement, systems and outcomes.
The AI dividend
AI creates an efficiency dividend.
The leadership question is simple:
Where does that dividend go?
Weak agencies use it to squeeze more output from the same tired team.
Smart agencies reinvest it into:
better thinking
stronger craft
training
experimentation
workflow improvement
proprietary systems
new revenue models
healthier culture
That reinvestment compounds.
If every efficiency gain just becomes “great, now do more”, the agency slowly turns into a content treadmill with better software.
Not exactly the dream.
Cutting people is the lazy answer
AI removes friction.
It does not remove the need for talented people.
The agencies that simply cut juniors because AI can write first drafts are creating tomorrow’s senior talent gap.
That is not strategy.
That is spreadsheet panic wearing a Patagonia gilet.
The smart move is role evolution.
Juniors become better researchers, editors, prompt-builders, curators and producers.
Producers become workflow architects.
Creatives become stronger directors of systems, taste and output.
Strategists spend less time gathering and more time interpreting.
Account teams become better at briefing, expectation-setting and explaining how AI-enhanced work gets made.
Leadership becomes capability architecture.
That is the staffing model worth building.
Small, sharp teams with strong systems will beat bloated teams stuck in process theatre.
But you still need excellent humans.
Possibly more than ever.
Pricing has to evolve
Hourly billing will not disappear overnight.
But relying on it blindly is dangerous.
As AI compresses delivery time, clients will increasingly ask:
“Why are we paying the same if it took less time?”
And they will have a point.
The answer is to stop selling time as the main product.
Better models include:
fixed-fee projects
strategy and concept sprints
value-banded retainers
retained advisory
AI-enabled workflows
licensed tools or systems
build-and-optimise models
capability programmes
The shift is from labour to leverage.
You are selling the outcome, the system, the thinking, the confidence and the momentum.
Not the number of hours it took to get there.
Still track time internally. You need to understand cost and margin.
Just stop making time the headline act.
The wrong metrics will hurt you
Some old agency metrics are starting to look a bit wobbly.
Headcount used to signal scale.
Now it can signal inefficiency.
Utilisation used to signal productivity.
Now it can mean nobody has time to improve workflows, learn properly or think beyond the next deadline.
Better metrics now include:
profit per employee
revert reduction
pitch velocity
margin quality
speed-to-market
workflow efficiency
strategic depth
client trust
capability density
That last one matters.
Capability density is how much value a team can create with the people, systems and tools it has.
A small team with high capability density can now punch horribly above its weight.
A large team with low capability density becomes expensive chaos.
The dangerous middle
Many agencies are currently stuck halfway through the shift.
They are faster than before.
But they have not changed pricing, positioning, workflows or commercial strategy.
So the agency gets quicker.
The client expects more.
The team gets busier.
Margins get thinner.
Everyone calls it transformation.
It is not transformation.
It is operational self-harm.
The practical move
You do not need to blow up your commercial model next week.
But you do need to start moving.
Start with this:
track where AI is reducing delivery time
identify which services are most exposed
move selected work to fixed fees
package repeatable offers
price around outcomes, not effort
build advisory or systems-based revenue
reinvest efficiency into capability
train the team to explain the value clearly
The biggest mistake is waiting until clients force the pricing conversation.
Better to lead it.
The real opportunity
AI should improve agency economics.
Not flatten them.
The agencies that win will not simply be faster.
They will be more leveraged, more focused, more useful and more commercially intelligent.
The old model rewarded scale.
The new model rewards capability density.
That is the shift.
And the agencies that understand it early will pull away fast.