Why does AI adoption stall in agencies?
Summary: Spark AI's 2026 agency benchmark found that 83% of agency staff say they are confident with AI, while only 15% of agencies have built it into how the business operates. That gap is the problem we get asked to solve more than any other.
Date: 07.08.2026
Author: Emma Wharton Love
Read length: 6 minutes
Spark AI is the AI performance partner that builds organisation-wide AI capability for creative agencies and brand teams. We take ambitious teams from fragmented experimentation to AI embedded across your people, processes and systems ā so it delivers real commercial impact.
AI adoption stalls in most agencies because individual fluency never becomes organisational capability. People get good at using AI. The business around them doesn't change.
āThe same sentence comes up on almost every call we take with agency leaders: people here are using AI, but it isn't sticking across the business.ā ā
It's a fair description of where most agencies are in 2026. Spark AI has run its agency benchmark every six months since 2024, and the picture is consistent. Agency teams have moved past the experimentation phase. They use AI regularly and they feel confident doing it. What we see when we look closely is still good but basic prompting ā people don't know what they don't know.
Why doesn't individual confidence add up to organisational capability?
Confidence using AI is a personal feeling and capability is a system, and one doesn't turn into the other on its own.ā
Spark AI's Spring 2026 agency benchmark found that 83% of agency staff say they're confident with AI, while only 15% of agencies are building AI into the way their business operates.
āThat means AI strategy, governance, commercial positioning and skills all remain stuck at the earliest stage of maturity, which creates a widening gap between what teams are doing with AI and what their organisations are set up to support.
āThe consequences run in both directions. Without governance, agencies are exposed to risk around IP, data and client confidentiality. Without strategy, they leave measurable gains in efficiency, creative capacity and competitive positioning on the table.ā ā
Someone who has worked out a good workflow has a personal habit. An agency that has documented that workflow, trained others on it and priced it into a proposal has a capability. Most agencies have plenty of the first and very little of the second.
What are the 15% doing differently?
ā They're asking leadership questions rather than technology ones.
ā āThe agencies making real progress ask three things:
How does AI reshape our business model?
How do we build repeatable workflows that deliver genuine value to clients, not just internal time savings?
How do we make sure the whole team is engaged, not just a handful of enthusiasts?
āNobody in that group is waiting for a better model to arrive. And the answers show up in two places you might not expect: client conversations, and valuations.ā ā
Why does AI governance keep getting skipped?
āBecause it's the least interesting item on any leadership agenda, right up until a client asks about it in a pitch.ā ā
Spark AI's 2026 AI in agencies benchmark The Spark Report found that 52% of agencies have no formal governance in place ā no whitelisted tools, no training on data handling. ā
Clients are paying closer attention now. They want to know you're handling their data responsibly. They also want to see what you can do with AI. Being ready for both is increasingly a determining factor in winning pitches, and in M&A conversations.
A defensible position doesn't need to be a forty-page policy. Three things get you most of the way:
A list of approved tools, and what each one is cleared for
A clear rule about which client data goes into which tool, and what never leaves your systems
A team who can answer confidently when a client asks how you're using AI in their work
āThat third one is the part agencies underestimate. A specific, confident answer from an account director does more for client trust than any policy document filed on the intranet.
What stops agencies going further?
āCapacity, not scepticism and not cost.
āSpark AI's research has consistently put capacity at the top of the barrier list across two years of benchmarking. People want to get better at this. They just can't find the time between client work and everything else. That holds even for experienced teams, which is why waiting for a quiet fortnight has never worked as a strategy.
āThe simplest intervention we know is to stop treating AI development as something that happens in the gaps. Block an hour a week, for yourself and your team, with no deliverable attached. Just space to try things and build familiarity. Protect it the way you'd protect a client meeting.
ā It sounds too small to matter. It's the most common difference we see between agencies that move and agencies that talk about moving.ā ā
What comes after the hour?ā ā
Structure. An hour a week creates the conditions; it doesn't create the system.
āThe agencies that convert individual experimentation into organisational capability do three more things: they put one named person in charge with protected time, they build a rhythm for sharing what works, and they capture what they learn so it doesn't leave when someone does. Our 90-day AI implementation plan sets out how to sequence all three.
Treating AI as the organisational change it is means training that goes beyond basic prompting, clear rules about what data goes into which tools, and a team that can talk about it credibly in front of a client. None of that requires a large budget. It does require someone to decide it matters.
Frequently asked questions
Isn't 83% confidence a good sign?
It's a better starting point than scepticism. The caution is that self-reported confidence and demonstrated capability tend to diverge ā most of the confident use Spark AI observes is competent basic prompting rather than sophisticated work. Confidence tells you people will engage. It doesn't tell you the business has changed.
What does "embedded" AI adoption mean?
In Spark AI's benchmark it means AI shows up in how the business operates rather than only in how individuals work: a stated AI strategy, governance a client could be shown, workflows documented well enough for a new joiner to follow, and a commercial position that reflects what AI has changed. Individual enthusiasm doesn't count towards it.
How many agencies have formal AI governance?
Spark AI's 2026 agency benchmark found that 52% of agencies have no formal governance in place, meaning no whitelisted tools and no training on data handling.
How do we find out where our agency sits?
The Spark Index scores your AI maturity against the Spark AI Maturity ModelĀ®, benchmarks you against 80+ agencies and brand teams, and gives you a 90-day plan. It runs as a team-wide survey plus leadership interviews. Book a call to find out what's involved.
Where can I find more AI insights like this?
Spark AI's benchmark research is published twice a year in The Spark Report. We also publish Spark Intelligence, a fortnightly newsletter for creative leaders, read by around 2,150 people across more than a thousand companies.
Spark AI ā AI training, consultancy and agent building for agencies and brands. Your AI performance partner. https://www.wearespark.ai/
