AI and the future of the agency
Preparing your agency for sale
Takeaways from our latest Founders Breakfast
To round off 2025 we gathered together 30 agency founders for our latest BLINK Breakfast. We focused on AI.
Tom Salmon from AgencybyAgency provided some thought-provoking insights on AI and automation before we opened up the discussion to the floor. Quite a range of disciplines were present, from branding and design to content creation and comms.
Here are the highlights and key takeaways from both the presentation and ensuing discussion.
1. This is structural, not cyclical
The data coming out of AgencybyAgency reinforces something many founders already feel: this isn’t a short-term wobble. Agencies are operating after more than a decade of economic distortion, in a market shaped by short-term ROI thinking and financial control.
A generation of marketing leaders has grown up optimising rather than building. That matters, because it changes how agency value is evaluated - and defended.
2. Agency growth still follows budget
AgencybyAgency’s data shows the clear relationship between where budgets move and where agencies grow. Growth is strongest in channels where clients still need external expertise.
It’s not about size. It’s about uncertainty. Agencies thrive in spaces where clients don’t yet know how to do that thing themselves.
3. Newer, harder work grows faster
When growth is measured by headcount rather than turnover (to strip out inflation), the fastest-growing agency subsectors are:
Digital transformation
Data and market analytics
Sustainability and social purpose
Amazon and retail media
Purpose- and sustainability-led agencies tend to be more deeply embedded, advising on organisational change rather than just running campaigns. That depth of integration is a leading indicator of resilience.
4. AI is already normalised - just not admitted
Around 75% of agencies are already using generative AI in their workflows. Almost everyone in the room confirmed this.
And yet only 23% of agencies mention it as part of their proposition.
That gap isn’t accidental. It reflects a deep tension: agencies are adopting AI quickly while being unsure how to talk about it without triggering pricing pressure or commoditisation.
5. The pricing model is the real problem
Some clients still insist on asking: “If you’re using AI, shouldn’t this be cheaper or faster?”
This exposes the fragility of time-and-materials pricing. If value is measured in hours, AI looks like a threat. If value is measured in outcomes, judgment, and decision quality, AI can turn into leverage.
6. Efficiency is what the economy demands - not what creates advantage
Most agencies are currently using AI in the first phase of transformation: streamlining and efficiency. That makes sense in a tight market.
But efficiency alone doesn’t build differentiation. The data suggests the real upside comes when AI is used to:
Expand what agencies can offer
Reimagine how clients experience value
Those phases take longer - possibly 2-3 years - but this is where defensibility lives.
7. AI lowers barriers - it doesn’t replace judgment
AI has dramatically reduced the cost of producing “good enough” work. That’s already leading some large brands to cut content production budgets.
But the data and discussion both point to the same conclusion: quality still replies heavily on human taste and judgment. AI accelerates output; it doesn’t validate it.
As ever, garbage in still produces garbage out - just faster.
8. The ‘creative middle’ is being hollowed out
High-fidelity AI output is eroding the space where agencies historically sat: the creative middle ground.
Clients are increasingly arriving with AI-generated concepts and asking agencies to execute or refine them. That raises uncomfortable questions about where conceptual value begins — and who pays for it.
Agencies are being pushed toward a clearer choice:
High-value strategic and conceptual leadership
Or high-volume execution at scale
Sitting in the middle is becoming increasingly difficult to sustain.
9. Briefs are getting worse, not better
Clients are using AI to write briefs - but those briefs often jump straight to solutions or focus on squeezing cost.
This makes the agency’s role in interrogating, reframing, and translating the brief more valuable, not less so. Poor inputs still produce poor outcomes, regardless of how advanced the tools are.
Frameworks that show value across inputs, outputs, outcomes, and long-term impact are becoming essential to move the conversation beyond hours and deliverables.
10. Talent isn’t disappearing - it’s being rewired
Job postings across the agency sector have slowed since 2022, signalling caution rather than collapse.
AI introduces a different opportunity: training systems on an agency’s own thinking, frameworks, and expertise. That allows junior talent to access senior-level judgment earlier - effectively a “grad scheme on steroids”.
The risk isn’t AI replacing people. The risk is using AI in ways that hollow out learning, progression, and long-term capability.
11. Legal reality is now a leadership Issue
Using client work as AI training data raises immediate legal and contractual questions. In most cases, the output belongs to the client, not the agency.
Very few agency terms currently reference AI usage, ownership, or protection. That gap won’t stay invisible for long. Agencies that address it early - ideally in partnership with clients - will be better positioned.
12. AI rewards industries that lean in
Research referenced in the session shows that industries most exposed to AI often see higher revenue per employee and faster wage growth.
AI tends to augment roles rather than eliminate them but only when organisations use it to elevate human contribution, not just reduce cost.
13. Optimisation and exploration need to coexist
A credible AI strategy allocates most effort to improving today’s business - but deliberately reserves time to explore what the business might become over the next 3-10 years.
That exploration works best when it’s led by people who understand the craft, not imposed as a top-down transformation programme.
14. Specialism still wins
The data continues to show that specialist agencies grow faster than generalists.
Clients still turn to agencies where something matters and they lack in-house capability - particularly in emerging or fast-moving disciplines. AI doesn’t change that. It sharpens it.
Final Reflection
AI isn’t eroding agency value. It’s exposing where value was never clearly articulated in the first place.
The agencies that will endure are the ones that stop defending hours, start defining impact, and use AI to scale their intelligence - and don’t apologise for adopting it.
AgencybyAgency provide comprehensive mapping of the UK agency sector: AgencybyAgency
