The hype around AI may be settling, but there’s still “a lot of skepticism” about what the technology can actually deliver, according to Northern agency leaders who recently came together to discuss its real-world impact on daily operations and client expectations.
AI was the hot topic at a roundtable titled ‘AI in Practice: From Hype to ROI – Exploring how agencies use AI, its impact on operations, and balancing human creativity’, hosted by Prolific North in partnership with Byteminds on 9 October at Manchester Museum.
From the importance of AI experimentation to achieving measurable ROI, balancing automation with creativity, and navigating shifting client expectations, the interactive event united agency leaders to share insights, challenges, and best practices.
Set against a backdrop of towering dinosaur fossils and ancient artefacts lining the halls of Manchester Museum, the setting felt apt as agency heads gathered in a conference room to debate a different kind of evolution: how AI is already transforming their digital agency operations and what it means for the future of the industry.
Attendees
- Mark Walton, founder, Byteminds
- Anton Permyakov, co-founder and CEO, Byteminds
- Leigh Gammons, CEO, Tangent
- Greg Sines, head of PHP, Reading Room
- Jonathan Healey, group technology director, IDHL
- Jonathan Seal, group strategy lead, Mando Group
- Christopher Burgess, CEO, CTI Digital
- Criss Richards, senior creative technologist (AI) McCann Worldgroup and IPG Global
- Joe Lyon, tech director, Velstar
- David Avis, client services director, AND Digital
- Lee Adams, founder, Cantarus
From AI fear to familiarity
The discussion, chaired by David Prior (Prolific North), kicked off with a welcome from Anton Permyakov, co-founder of Byteminds, followed by an honest reflection on how agencies have woven AI into everyday operations over the past six months.
From junior developers to creatives and marketers, many agency heads noted the initial anxiety and “levels of fear” some staff felt when faced with AI, either with concerns it would replace their jobs or that they wouldn’t pick up tools quickly enough.
“Less than 20% of our developers across the business were using AI,” said Christopher Burgess (CTI Digital). To shift perceptions, the agency launched innovation workshops and internal hackathons, encouraging teams to experiment. By June, internal surveys showed 78% of staff were using AI in some capacity, signalling a “cultural shift” as teams began viewing it as a tool, not a threat.
Similarly, Jonathan Seal (Mando Group) said his agency set KPIs for heads of department to experiment with AI, linking it to innovation workshops and employee satisfaction scores. While Criss Richards (McCann Worldgroup) said those practical workshops “changed the conversation completely” when staff saw AI tools as an opportunity to automate more tedious tasks.
Others, like David Avis (AND Digital), emphasised the importance of cross-departmental adoption. “We’ve seen the biggest wins where teams collaborate around a shared AI process rather than each team experimenting in isolation.”
For Jonathan Healey (IDHL), he would rather “keep pace with AI” as we may be underestimating its long-term impact. As it’s a new tech, there is always going to be a “learning curve”.
Moving from hype to ROI
When the conversation turned to whether AI tools or practices are really generating value yet, Leigh Gammons (Tangent) voiced concerns: “If we don’t change the fundamentals of how we work, we struggle to turn AI into real efficiency. Old process, new tools rarely makes us faster and cheaper.”
While the majority of agency leaders agreed that AI has moved beyond the hype stage and is beginning to demonstrate value for some, others would argue its results and readiness are “still up for debate” as it can be “flawed” for certain tasks.
Gammons (Tangent) noted how effective AI can be in development, sharing that two separate teams worked on the same project — one using AI and one not — with the AI-assisted team completing their work “16% faster”. But he questioned whether it has truly changed business models, observing that while AI can help speed up processes, it isn’t a dramatic transformation yet.
For other leaders, value is surfacing in other ways thanks to AI tools that automatically transcribe meetings and cut down wasted time.
Human creativity vs automation
As one leader points to a “level of skepticism or nervousness” about AI’s long-term role, Prior (Prolific North) asked attendees whether greater automation and less human involvement were inevitable.
Greg Sines (Reading Room) admitted to being “a bit underwhelmed at the moment,” noting that “there’s still a role for people”.
Confident there’s “significant work for us for years to come”, Jonathan Seal (Mando Group) explained how clients “have no way of keeping up with all the changes, capabilities, and prompts,” so from his perspective, agencies will evolve and shift where they spend their time.
Lee Adams (Cantarus) offered a metaphor from 1950s science fiction: “Humanity reached a point where the most valuable people were those who could come up with new questions. Maybe a lot of innovation will come from asking the right questions.”
If you were to fully automate everything, the output would be “garbage,” said Richards (McCann), adding that there will “always be a role for people” and clients “won’t pay for full automation”.
“Anyone can create something really quickly, it doesn’t mean it’s any good!” he added.
The discussion turned to how AI works with creativity, especially when it comes to tapping into nuance or tone of voice (TOV). Although “it is possible” to train AI agents to have a TOV, Seal (Mando Group) asked if agencies are happy giving it “that much control”.
Adams (Cantarus) and Burgess (CTI) questioned the nature of AI “intelligence” itself. Burgess (CTI) explained: “You need to have human interaction. If you prompt [AI tools] properly, you can get a better response.”
Evolving client expectations
When asked how AI is affecting client expectations, Lyon (Velstar) said most clients just want to feel “reassured” that an agency is embracing it and understands it.
Burgess (CTI) echoed that sentiment, while Richards (McCann) explained that unless AI features are adding something of value, it isn’t having an impact on expectations.
By contrast, Seal (Mando Group) reported that for enterprise clients “AI is front and centre” as these organisations are pursuing “major marketing transformation” and need help understanding what’s possible. And for clients in regulated industries, use of AI needs to be “bomb proof”.
According to Mark Walton (Byteminds), there are “impressive” tools and new workflows and from a development perspective, it can boost efficiency. But he warned about how this shifts where clients spend their money, sparking a conversation in the room about the need for billing models to potentially evolve towards “value” rather than rates.
Looking ahead, Adams (Cantarus) and other leaders in the room argued agencies must become trusted partners and adapt their service offerings to support their clients: “Non-tech businesses are going to struggle to keep on top of AI. Companies like us can help them use it and avoid the pitfalls.”
Career paths in agencies are already shifting and may lean into specialisms, explained Lyon (Velstar). Yet Richards (McCann Worldgroup) said roles are already changing as people are now in more “generalist” roles.
The roundtable wrapped up with the shared view that AI should be embraced by agencies, but making it truly valuable requires careful integration, hands-on oversight, and on-going experimentation, all while keeping pace with changing client expectations.
