The Spark Report – AI in agencies: Who's benefitting? Winter 2026/27
90% of agency staff now use AI. Only 6% of agencies have turned it into new revenue.
Key findings
Six findings from this edition
AI is now part of everyday agency work, and the personal gains are clear. Most agencies have not yet turned them into shared capability or new revenue.
Swipe for all six findings →
The big picture
Everyone is saving time. Few agencies are capturing the value.
Each chapter follows the value one step further – from the individual to the organisation, the agency and its future talent.
AI use is almost universal. New revenue from it is rare.
Five signposts from this report, %
See the data
| Staff using AI | 90% |
| Users saving time | 93% |
| Saved time → new work | 7% |
| Redesigned workflows | 11% |
| New AI revenue model | 6% |
Sources: Spark AI staff survey (first two bars) and Spark AI Maturity Score (last three), Winter 2026/27. Bases differ – read as signposts, not a funnel.
The Spark AI Maturity Model®
The average agency scores 1.8 out of 4
Nearly a third of agencies at Stage 2 or above are still at Stage 1 in at least one dimension.
Increasing maturity →
Stage 1Experimentation
- Vision & Strategy
- No AI strategy exists
- People, Skills & Culture
- Self-taught, uneven skills
- Data & Tools
- Personal tools, no data structure
- Governance & Accountability
- No policy, down to the individual
Stage 2Adoption
- Vision & Strategy
- Goals agreed, not written down
- People, Skills & Culture
- Shared training, sharing becomes routine
- Data & Tools
- Company accounts, single source of truth
- Governance & Accountability
- Policy exists, adherence unchecked
Stage 3Optimisation
- Vision & Strategy
- Documented strategy, commercial model shifting
- People, Skills & Culture
- Role-specific development, time protected
- Data & Tools
- Configured assistants, structured knowledge
- Governance & Accountability
- Policy checked, formalised in client contracts
Stage 4Transformation
- Vision & Strategy
- AI is a competitive differentiator
- People, Skills & Culture
- AI-native, continuous by default
- Data & Tools
- Connected systems, agency second brain
- Governance & Accountability
- Proactive governance, a commercial asset
Swipe to see all four stages →
Scored across Vision & Strategy, People, Skills & Culture, Data & Tools, and Governance & Accountability. Source: Spark AI Maturity Score, Winter 2026/27, n=[n].
Where does your agency sit?
Free, five minutes, with a tailored report on what to focus on next.
01 – People, Skills & Culture
90% of staff use AI. 93% of them say it saves time.
AI use is no longer limited to a small group of enthusiasts. 90% of the people we surveyed use AI at work, and 93% of those users say it saves them time.
Use has doubled since 2025
45% of agency staff we surveyed used AI in 2025. By spring 2026 it was 84%. Now it's 90%.
For the first time in our sample, people outside development roles reported using AI for coding and digital building. OpenAI's analysis of more than 800,000 work-related ChatGPT messages found 16.8% involved tasks associated with another occupation. AI is expanding what people can take on beyond their role.
AI use has doubled since 2025
% of agency staff using AI at work
See the data
| 2025 | 45% |
| Spring 2026 | 84% |
| Winter 2026/27 | 90% |
Source: Spark AI pre-programme staff survey, 2025 to Winter 2026/27. Spring figure as published. Base: all respondents, n=[n].
The personal productivity gain is growing
17% now save five or more hours a week, up from 5% in 2025. Only 6% save no time at all, down from 55%.
As AI use grows, the savings build across more of the working week. Models can now produce usable documents and presentations, so each use covers more of the task.
Weekly hours reclaimed by AI
% of AI users, three survey waves
See the data
| No time saved – Feb–Oct 2025 | 55% |
| No time saved – Oct 2025–Apr 2026 | 11% |
| No time saved – Winter 2026/27 | 6% |
| 1–3 hours – Feb–Oct 2025 | 34% |
| 1–3 hours – Oct 2025–Apr 2026 | 55% |
| 1–3 hours – Winter 2026/27 | 49% |
| 3–5 hours – Feb–Oct 2025 | 6% |
| 3–5 hours – Oct 2025–Apr 2026 | 28% |
| 3–5 hours – Winter 2026/27 | 28% |
| 5+ hours – Feb–Oct 2025 | 5% |
| 5+ hours – Oct 2025–Apr 2026 | 6% |
| 5+ hours – Winter 2026/27 | 17% |
Source: Spark AI pre-programme staff survey. Base: respondents who use AI, n=[n].
Who's benefitting?
The individual
People are completing tasks faster, reclaiming meaningful time and taking on a wider range of work. But activity tells us little about depth of adoption. Maturity is about what the organisation builds around that use – and whether personal gains become shared capability.
How to find out where your team's saved hours are going – and a simple way to track them.
02 – People, Skills & Culture, Governance & Accountability
Only 7% are using saved time to unlock new kinds of work
AI is saving people time. The question is whether those individual gains become shared practices the business can capture.
Where agencies sit – People, Skills & Culture
- Stage 1 · ExperimentationSelf-taught, uneven skills35%
- Stage 2 · AdoptionShared training, sharing becomes routine41%
- Stage 3 · OptimisationRole-specific development, time protected18%
- Stage 4 · TransformationAI-native, continuous by default6%
Source: Spark AI Maturity Score, Winter 2026/27. Highlighted: where most agencies sit.
Where the saved time goes
47% said AI helped them produce more in the same time. That's an efficiency gain – but is the work better, or only faster?
43% said saved time was absorbed or hard to account for. The agency can't choose where to put that capacity if it can't see where it goes. Only 7% said it had enabled new kinds of work. Chapter 3 looks at how redesigning a complete workflow could make that happen more often.
Only 7% are using saved time for new kinds of work
% of respondents – more than one answer allowed
See the data
| More output, same time | 47% |
| Absorbed or hard to account for | 43% |
| Better-quality work | 39% |
| Less workload pressure | 24% |
| New kinds of work | 7% |
Source: Spark AI pre-programme staff survey, Winter 2026/27, n=110. Respondents could select more than one outcome, so percentages don't total 100%.
Better work and less pressure pay off
39% said AI helped them produce better-quality work – sharper thinking and a higher standard of delivery that clients feel directly.
24% said AI reduced their workload pressure. In an industry where burnout is the norm, that has a commercial pay-off.
"When people work sensible hours, they stay. That means less time and money lost to recruiting and onboarding, less knowledge walking out of the door, and clients who keep a team that knows their business. AI gives agencies a way to get there – if they protect the time it frees up."Jo Higgs, Event Concept
A third of agency knowledge still lives in people's heads
Shared practice depends on shared knowledge. When the way an agency works stays in people's heads, successful AI use is hard to repeat.
Documenting the process and what good looks like gives the team – and its AI systems – a consistent reference point.
How agency knowledge is held
% of agencies
See the data
| Informal, in people's heads | 35% |
| Agreed source for key documents | 40% |
| Structured and current for AI to use | 16% |
| Judgement encoded into AI systems | 8.5% |
Source: Spark AI Maturity Score, Winter 2026/27, n=117. Single answer.
Two-thirds of agencies don't check how AI is used
38% have an AI policy but don't check it's followed. A further 29% leave AI use unmonitored. Only 33% actively apply or review their rules.
For most agencies, governance has reached Stage 2 – Adoption. It isn't yet part of everyday practice.
How AI governance is applied
% of agencies
See the data
| How AI governance is applied – Unmonitored | 29% |
| How AI governance is applied – Policy, not checked | 38% |
| How AI governance is applied – Actively applied or reviewed | 33% |
Source: Spark AI Maturity Score, Winter 2026/27, n=93.
Where agencies sit – Governance & Accountability
- Stage 1 · ExperimentationNo policy, down to the individual28%
- Stage 2 · AdoptionPolicy exists, adherence unchecked43%
- Stage 3 · OptimisationPolicy checked, formalised in client contracts23%
- Stage 4 · TransformationProactive governance, a commercial asset6%
Source: Spark AI Maturity Score, Winter 2026/27. Highlighted: where most agencies sit.
still rely entirely on self-directed AI learning
have a named AI lead or taskforce driving adoption
have no named owner to drive AI adoption
General training gives the team a shared baseline and common language. Role-specific development helps people apply AI to their own work, and continuous learning keeps that capability current.
A named owner creates accountability for spotting what works and turning it into shared practice. Without one, nobody is formally responsible for moving good methods beyond the people who developed them.
Swipe for more →
Source: Spark AI Maturity Score, Winter 2026/27. Ownership base n=47.
Who's benefitting?
The individual first. The organisation when practice is shared.
Shared standards turn one person's effective use of AI into a way of working the wider team can repeat. That's the shift from personal productivity to organisation-wide capability.
A prompt for building an SOP skill, a plan for an AI away day, and Ali Hanan of Creative Equals on tackling bias.
03 – Data & Tools
Only 11% have redesigned the workflow
Once an agency has a shared way of working, it can redesign the workflow around those standards. This is how shared practice becomes organisational capacity, new capability and value for clients.
Where agencies sit – Data & Tools
- Stage 1 · ExperimentationPersonal tools, no data structure32%
- Stage 2 · AdoptionCompany accounts, single source of truth47%
- Stage 3 · OptimisationConfigured assistants, structured knowledge17%
- Stage 4 · TransformationConnected systems, agency second brain4%
Source: Spark AI Maturity Score, Winter 2026/27. Highlighted: where most agencies sit.
Most AI use is still bolted on
24% use AI for productivity tasks and 46% to improve client work within the existing process. Only 11% have end-to-end workflows with defined roles for people and AI.
A further 19% have configured AI assistants. That can make one part of the process more consistent, but it doesn't necessarily change how the whole job gets done.
Only 11% have redesigned the workflow end to end
How AI is used within workflows, % of agencies
See the data
| Productivity tasks | 24% |
| Improving client work in the existing process | 46% |
| Configured AI assistants | 19% |
| End-to-end redesigned workflows | 11% |
Source: Spark AI Maturity Score, Winter 2026/27, n=[n].
Redesign changes how the work flows
AI can speed up every stage of an existing workflow while leaving the process itself unchanged. People have traditionally worked through a job one step at a time. AI can be configured to carry out several connected steps at once – which changes how the work flows and makes room for things that used to take too long, offering new possibilities to clients.
Redesign also creates space to look at how information moves between people and what the next person actually needs. It can cut cognitive load and context switching, while keeping human judgement at the centre.
Agents have moved the opportunity forward
Agents have been a defining development since the Spark Spring Report. An agent is a narrowly defined job that runs on a schedule or a trigger, reaches into your systems and leaves the result somewhere useful.
They can be pointed at repetitive, manual tasks. They aren't workflow redesign on their own, but they've expanded what agencies can redesign and what those workflows can produce.
Agency size changes the shape of AI maturity
Micro and small agencies have the most balanced profiles. The gap between their strongest and weakest dimensions is 0.3 and 0.2 points. In medium agencies it widens to 0.7.
Larger agencies often need governance and formal structures before tools can roll out across teams. Smaller agencies have fewer layers to coordinate, so they can test and adapt more readily. That agility is a new advantage, but progress still depends on all four dimensions moving together.
Workflow redesign brings all four into the same piece of work. The workflow becomes the unit of change, moving the organisation through the maturity model as a whole.
Bigger agencies can be strong in one area and weak in another
Gap between each size of agency's strongest and weakest area of AI maturity, in points
See the data
| Micro – Fairly even across all four areas | 0.3 pts |
| Small – Fairly even across all four areas | 0.2 pts |
| Medium – Strong on governance (2.3), weak on data & tools (1.6) | 0.7 pts |
| Large – Strong on governance (2.8), weak on strategy (2.0) | 0.8 pts |
Source: Spark AI Maturity Score, Winter 2026/27. Large-agency sample is small – treat as directional.
Who's benefitting?
The organisation, when the work changes
Individuals benefit first when AI makes their tasks faster. The organisation benefits when AI changes how the work gets done and creates better work, new services or new value for clients.
How to redesign your first workflow, and a checklist for building a competitor-watch agent.
04 – Vision & Strategy
Only 6% have created new AI-enabled revenue models
Redesigning the workflow changes what an agency can deliver. Strategy defines the commercial aim. Pricing determines who captures the benefit.
Where agencies sit – Vision & Strategy
- Stage 1 · ExperimentationNo AI strategy exists32%
- Stage 2 · AdoptionGoals agreed, not written down44%
- Stage 3 · OptimisationDocumented strategy, commercial model shifting21%
- Stage 4 · TransformationAI is a competitive differentiator3%
Source: Spark AI Maturity Score, Winter 2026/27. Highlighted: where most agencies sit.
Commercial change needs direction
65% of agencies don't have a documented AI strategy. Most recognise AI's potential, but only 35% have written down what they want it to achieve.
Without that direction, it's harder to connect changes in how work is done to the value the agency wants to create and capture.
Two-thirds of agencies don't have a documented AI strategy
% of agencies
See the data
| How well defined is the AI strategy? – No formal strategy | 27% |
| How well defined is the AI strategy? – Agreed leadership goals | 38% |
| How well defined is the AI strategy? – Documented, linked to goals | 24% |
| How well defined is the AI strategy? – Tracks milestones | 11% |
Source: Spark AI Maturity Score, Winter 2026/27, n=[n].
Pricing determines who captures the benefit
Output-based pricing is now the largest single group, at 43%. But 32% still price purely on time, and only 6% have reached a new AI-enabled revenue model.
Only 6% have built a new AI-enabled revenue model
How agencies price and sell their work, % of agencies
See the data
| Output or deliverable | 43% |
| Input or time-based | 32% |
| Value-based | 19% |
| New AI-enabled revenue model | 6% |
Source: Spark AI Maturity Score, Winter 2026/27, n=[n].
"The most exciting conversations are about things clients have never been able to do before. That's where AI opens up new budgets, rather than squeezing the ones we already have."[Name], [job title], IPA
The client conversation hasn't caught up
58% have no formal client position on AI. Pricing changes are hard to sustain without a clear conversation about how AI is used and what has improved for the client.
As AI takes on more repetitive groundwork, teams have more room for client relationships and decisions that need human judgement. That makes space for a better conversation about the value of the work, rather than reducing AI to time saved.
How agencies discuss AI with clients
% of agencies
See the data
| How is AI discussed with clients? – Not acknowledged | 15% |
| How is AI discussed with clients? – Discussed informally | 43% |
| How is AI discussed with clients? – In contracts or terms | 24% |
| How is AI discussed with clients? – Part of why clients choose them | 18% |
Source: Spark AI Maturity Score, Winter 2026/27, n=[n].
Who's benefitting?
The agency – when strategy, pricing and the client proposition reflect how the work has changed
Without that shift, greater efficiency may simply reduce billable time or pass the benefit straight to the client.
An exercise to show clients what AI makes possible and make the case for value pricing, plus the EU AI Act checklist for agreeing AI responsibilities with clients.
05 – People, Skills & Culture
Under-25s in agencies fell by 19.2% in 2025
Workflow redesign can create new commercial opportunities. Our hypothesis is that it can also create a stronger role for junior talent. As AI takes on early-stage tasks, juniors can take greater ownership of complete outputs, with senior judgement brought in while decisions are being made.
Where agencies sit – People, Skills & Culture
- Stage 1 · ExperimentationSelf-taught, uneven skills35%
- Stage 2 · AdoptionShared training, sharing becomes routine41%
- Stage 3 · OptimisationRole-specific development, time protected18%
- Stage 4 · TransformationAI-native, continuous by default6%
Source: Spark AI Maturity Score, Winter 2026/27. Highlighted: where most agencies sit.
Entry-level routes are narrowing
employees aged 25 and under across IPA agencies
vacancies across all levels
agencies employing graduates or apprentices
AI is one factor within wider commercial pressure, but junior roles are especially exposed because many of their traditional tasks can now be done by AI.
Source: IPA Agency Census 2025.
The same pressure shows up beyond UK agencies. Employment among 22- to 25-year-olds in AI-exposed US occupations is 19% below where it would be if it had kept pace with less-exposed roles – driven mainly by reduced hiring (Stanford Digital Economy Lab).
Workflow redesign can create stronger junior roles.
Redesigning the workflow changes how a job is divided and how it moves between people. As AI takes on some early tasks, juniors could move beyond owning individual steps to taking responsibility for a complete output.
A senior works with them as decisions are made, helping them develop the judgement to take on more of the process over time. Those decisions can be turned into a reusable AI Skill, so the next time a junior does similar work, the guidance is there.
For agencies, this is a way to develop future talent while changing how work gets done. Cutting junior hiring may reduce costs today, but it leaves fewer people gaining the experience agencies will need later.
Who's benefitting?
Junior talent and the agency
Juniors gain broader responsibility and a new route to developing judgement. Agencies gain adaptable talent and capture knowledge the wider team can share.
A discussion guide to choose a piece of work a junior could own, and how to turn the decisions into a reusable AI Skill.
How does your agency compare?
Take the free AI Maturity Score to see your stage on each dimension, benchmarked against the agencies in this report.
About the research
How was the Spark Report research done?
Two sources. The Spark AI pre-programme staff survey (n=110 this edition) shows how agency staff use AI. The Spark AI Maturity Score (up to n=117 agencies) shows how agencies are organised around AI. Some questions were added during collection, so each chart shows its own base. Written by Asta Vallis, Emma Wharton and Jules Love.
What is the Spark Report?
The Spark Report is Spark AI's twice-yearly research into how creative agencies and brand teams are adopting AI, and what is driving real value.
How many agency staff use AI?
90% of agency staff use AI at work, up from 45% in 2025 (Spark AI pre-programme staff survey, Winter 2026/27).
What is the Spark AI Maturity Model®?
The Spark AI Maturity Model® is a four-stage model – Experimentation, Adoption, Optimisation, Transformation – scored across Vision & Strategy, People, Skills & Culture, Data & Tools, and Governance & Accountability. The average agency scores 1.8 out of 4.
Is the full report free?
Yes. Fill in the short form on this page and we'll email you the PDF, including a practical plan for every chapter.
Get the full report – what to do about it
This page shows what we found. The free PDF shows what to do next.
- A practical plan for every chapter
- The EU AI Act checklist for agencies and brands
- Pricing models agencies are moving to
- How to redesign your first workflow
- Contributions from industry leaders
The Spark Report – Winter 2026/27
Change the work to capture the value
Free PDF · [n] pages · a practical plan for every chapter
Fill in the short form below and we'll email it to you straight away ↓
Add form from The Spark Report page