AI dominated the agenda as agency leaders met in Leeds for Prolific North’s From Growth to Gold summit, where experts warned the industry is only “2% prepared” for what’s coming.
Backed by headline sponsors KPMG and Twisted Loop, the 2025 Leeds edition cut through the usual conference gloss with a mandate to interrogate what’s really happening in M&A, examine how AI is reshaping value, and spell out what future-ready agencies must do to stay investable.
Chairing the day, Richard Gregory, co-founder of Agency NXD and a veteran of multiple agency growth cycles, set out the refreshed format, which included a fast run of morning keynotes and panels followed by “not-quite-round” roundtables designed to push delegates into frank, practical discussion rather than passive listening.
READ MORE: The Prolific North Top 50 Digital Agencies 2025
KPMG partner Ben Taylor set the economic context. He leads the firm’s technology M&A focus in the region, with a particular eye on digital media and transformation services.
Deal flow in tech is down around 10% over the last 12 months, he told delegates, and more like 20% in the marketing and digital services space. The picture on spend is brighter with ad, marketing and digital budgets becoming more “stabilised” and showing consistent growth over recent quarters. But when it comes to buying and investing in agencies, acquirers are being far more selective.
Where once buyers were happy to “buy capacity”, Ben said the conversation now starts with a much harder questions about what makes you unique and different? That theme of distinctiveness – who you are for, and why you exist – would echo through the rest of the day.
The AI wave is coming – and agencies aren’t ready
If Ben framed the financial reality, Twisted Loop’s Jonathan Summers zoomed out to the technology curve. With more than 30 years in data and AI systems, he argued we are “conservatively 18 to 24 months away from 10x productivity” as human teams and AI systems knit together. By 2026, he predicted, a huge increase in data centre capacity will deliver roughly a thousand times more compute than went into GPT-3, much of it focused on agentic AI – “agents doing AI research.”
His short-term prediction was even bolder. By the middle of next year we should expect AI agents running autonomously for eight hours a day at very low cost, he said. “The cost of thought is now tending to almost zero,” he said. By the end of 2025, he expects to see agents with human-level expertise in certain sectors.
Twisted Loop has recently surveyed marketing agencies on their use of AI. Jonathan shared a few headline stats:
- Around 90% of agency professionals say they’re using AI.
- Only 2% feel prepared for what’s coming next.
- 70% put internal efficiencies in their top three use cases.
- 83% haven’t built any bespoke tooling inside their organisations.
- 89% feel positive about AI in their industry, but 66% fear over-reliance.
For Jonathan, that adds up to an industry that is “highly optimistic, but not blindly so” – and one where the gap between strategic intent and operational execution “is not closing.” Agencies want to do it; they’re just not doing it in a structured, deliberate way.
His prescription that everyone in the business needs higher AI fluency; every agency needs an operational blueprint for how AI will be used; and leaders should be building and experimenting, not just talking. The rest of the programme was effectively a live exploration of how that might look in practice.
“Could we do it again – and bigger?”
The opening keynote came from Connective3 founder and CEO Tim Grice, who was interviewed on stage by Richard.
Tim traced his journey from “messing around trying to be a lawyer” to joining a tiny production agency in 2009 that would become Branded3. When he arrived, it was effectively a two or three-person SEO shop. Over the next few years the team grew to around 60 before being sold.
Crucially, it wasn’t his business that exited. That itch – to build something and realise the value personally – was one of the reasons he came back for a second run, founding Connective3.
GET ALL THE LATEST HEADLINES IN YOUR INBOX HERE
“I wanted to have a go at building that value, doing it for myself,” he explained. There was also a nagging question. Had Branded3’s success been down to timing, or could he and his colleagues repeat it – “do it again, bigger and better”? And fundamentally, he said, he just loves agency life and the chance to create new roles and “change people’s lives.”
Second time around, the backdrop is very different. Hybrid and flexible working are now the norm, and he was candid about the cultural challenges.
“Silos were supposed to be a thing of the past, but actually when you’re working remotely they’re more often than not something you still need to deal with,” he said. Productivity isn’t the issue; communication and culture are. At the same time, clients are demanding more integrated responses. The days of narrow, channel-specific briefs have given way to “connected, cross-channel” mandates that require broader marketing and digital skills, he said.
On international expansion, Connective3 made an early decision not to chase the US – at least not yet. Post-lockdown, they began to see strong inbound interest from Europe instead. Around 10% of revenue now comes directly from Germany, supported by a German team, with additional work from France, Spain and Italy. Tim has spent much of the year on the continent and sees a clear opportunity: “I think we’re slightly ahead in terms of our thinking and the work that we’re doing,” he said, and European clients are outsourcing to the UK to tap into that.
From sceptic to “fully converted”: betting big on AI
Like many in the room, Grice admitted he was sceptical about AI as recently as 12 months ago. That’s changed dramatically in the last six to eight months.
Connective3 has invested just over £250,000 and built a five-person innovation and AI team, combining prompt engineers, full-stack developers and data scientists. They’ve launched a platform that houses all their technical tools, stitched together and controlled through an internal chatbot. Rather than logging into separate utilities or filling out forms, staff can simply ask the bot to “check this URL” or run an audit; AI does the “heavy lifting” and feeds back insight for humans to interpret and turn into strategy.
AI is also embedded in their own build process. Tools are now automatically tested by AI, which generates bug reports for the team. Once a human has sense-checked and approved fixes, AI is used again to implement them. With that setup, a small team has been able to ship new tools at what Tim described as “ridiculous speed”. In one case, a developer built a new tool “while he was watching Netflix” at home.
For clients, that shift is already changing what they buy. Technical auditing work they once charged for is now commoditised. “A client can get an audit by going online,” Tim said. What they’re willing to pay for is what to do with it – strategy, prioritisation and creative thinking. His view of the future is fewer people doing process, with more people doing thinking, creativity and higher-value consultancy.
Connective3 is currently in market raising minority investment to accelerate acquisitions in creativity, social and influencer work and to push further into Europe. He was open about how draining that process can be. It’s “exhausting,” he said. Distracting, emotional, and full of spreadsheets too, he joked. His advice for other founders was to have “no expectations” on timescales or valuations, get ready for a rollercoaster – and prepare to have your “baby” poked and prodded by investors who don’t automatically share your view of its brilliance.
Agency models under pressure
If the keynote focused on one founder’s journey, the first panel zoomed out to the wider agency model. Andrew Durkin from Answer Digital described a 25-year-old consultancy that has recently transitioned to an employee ownership trust. Around 60% of its revenue comes from the NHS and 40% from wealth management, insurance and private-equity-backed businesses. Their commercial model mixes fixed-fee discovery “loss leaders” – typically four-week engagements to assess the state of a client’s platform – with longer programmes sold on time and materials, structured through rolling statements of work.
Visualsoft CEO Chris Fletcher shared how the 27-year-old ecommerce business is mid-transition from a proprietary platform to a Shopify-first, agency-led model. Pre-Covid, Visualsoft sold a simple performance deal – a single monthly fee, often pegged at around 3% of turnover, in return for platform and full-funnel marketing. That model worked during boom times but broke down as markets softened and clients demanded more transparency. Now, project-based Shopify builds sit alongside separate retainers for marketing services, forcing a wholesale reset of the company’s structure, leadership and P&L.
Fletcher was frank about price pressure. Builds that once commanded £200,000 can now be bought for £5,000–£10,000 from small shops born in the post-Covid boom. Visualsoft is using AI to streamline discovery, scoping and development – improving margins while keeping price points competitive – but, he stressed, you still “always have people there” on top. Tools get you so far; humans still need to own the work.
For Ginny Nicholls, founder of Interim Digital, the answer has been to combine agency-style thinking with the flexibility of a vetted freelance collective. Interim charges on straightforward day rates but sells outcomes and agility – allowing brands and agencies to plug in senior specialists without the overhead of a full-time hire. AI is already deeply embedded in their processes, from sales outreach to removing grunt work for freelancers, freeing them to focus on higher-value tasks and, in some cases, to serve more clients.
Bringing a data lens to the discussion, Agency by Agency co-founder Tom Salmon drew on a database of 25,500 UK agencies built to map specialisms, growth and competition. The number of agencies in the UK has doubled between 2015 and 2025, he noted, making this “one of the most competitive and advanced” markets in the world.
The agencies that perform best, Salmon said, are those that know who they are for, when they are most useful, and how they deliver value – and can articulate that beyond generic claims about “our people” or a list of services. Awards matter, but with roughly a fifth of agencies award-winning, they’re nowhere near enough to differentiate on their own.
On AI, his numbers echoed Twisted Loop’s. Around three-quarters of agencies are already using generative AI somewhere in their workflow, he said, but only around a quarter talk about AI or automation in their outward-facing proposition. For many, that hesitancy is tied to commercial models still built on time and materials.
On in-housing, Salmon said their data doesn’t support the idea of an unprecedented wave; instead, some skills are simply easier to bring inside than others. Where an agency owns a genuinely scarce specialist skillset it’s more insulated.
AI & growth: beyond cost-cutting
The final panel of the morning put AI centre-stage, with Tangent CEO Leigh Gammons and NAITIV founder and AI-first growth consultant Anthony Diver in conversation.
Leigh outlined how Tangent has built on his previous experience setting up an AI academy within a large network agency, this time embedding an AI division inside a mid-sized independent. He sees two particularly strong near-term use cases: customer support, where 50–60% of queries can be resolved without a human if the experience is well designed; and marketplaces, where AI-driven matching and decision-support can transform buyer–seller journeys.
But, he warned, getting those experiences right means serious attention to UX and guardrails. He cited one project where a gym’s chatbot cheerfully hallucinated its way into promising “beautiful spa facilities” that didn’t exist – a reminder that LLMs need constraint as well as creativity.
Anthony, who has spent the last two years exploring what an “AI-first” agency could look like after exiting his previous group, sees the frontline challenge as educational. Some brands and in-house teams are still wary or sceptical; others have spent 24 hours with the latest model and want to rebuild everything immediately. His argument is that AI projects must start with a clear problem and value proposition, not a vague ambition to “reduce costs”.
He pointed to examples beyond the usual brand names – such as Barking and Dagenham Council’s use of AI to improve outcomes around homelessness – as a reminder that the most meaningful use cases don’t always sit in glamorous sectors.
For both panellists, the commercial conversation around AI needs to get more honest. Many SMEs are looking at AI projects in the £50,000–£250,000 range with “low budget, high expectation” dynamics, Anthony said, while even giants have burned through hundreds of millions on initiatives that haven’t delivered consistent ROI. Clear staging – proving safety and basic functionality before promising step-change revenue or margin gains – is essential.
Leigh expects the next 12–18 months to bring not just more production deployments, but also the first high-profile failures of fully agentic systems – and with them, tougher questions about accountability and regulation. When something goes wrong, he asked, who is at fault: the developer, the vendor, the brand, the CIO?
Both also raised talent implications. Entry-level “spreadsheet monkey” work is already being eroded, even as AI creates new roles like prompt engineers and product-minded generalists. Diver argued that “tastemakers” – people with a strong sense of what good looks like in writing, design or experience, backed by data – will become more important, acting as quality control and ethical check on automated output.
Roundtables: from theory to action
After lunch, the room broke into smaller groups for themed roundtables chaired by KPMG, Twisted Loop, AgencyxAgency, Agency NXD and Prolific North.
Across sessions on agency M&A, AI and tech, accelerating sales, NEDs and growth, and global expansion, attendees tackled practical questions: how the current climate is shifting valuations and appetites; what really makes an agency attractive to buyers beyond revenue; when it becomes essential to bring in a Non-Executive Director; which go-to-market models work best when testing the US or Middle East; and how to build credible AI capability that actually strengthens valuation rather than just cutting costs.
Live polling from the stage made it clear that the room was split on AI – from hardened sceptics to those who fear for their business model within two years. But by the end of the day, a few threads had emerged: the importance of being distinct and data-literate; the need to bake AI into operations, not treat it as a bolt-on; and the reality that growth, funding and exits are marathons, not sprints.
