How SampsonMay turned individual AI use into a shared way of working
SampsonMay builds brands and digital platforms for global corporate clients, bringing strategy, creativity and technology together. After reading Jules’ book Shift: AI for Agencies, the leadership team paused their own AI rollout and joined the Spark AI Accelerator to get the next stage right.
The challenge
SampsonMay had already bought the tools, with Copilot across the agency and ChatGPT for the roles that needed it. What they had yet to put in place was an agreed way of using them. All five leaders described the same gaps: strategy, policy, structured training and ownership were all still to be defined, and usage across the wider team largely came down to individual preference. Many of their clients are in regulated sectors where data and intellectual property come under close scrutiny, which made settling that agreement particularly pressing.
How Spark AI helped
Spark AI ran our three-month AI Accelerator programme across the whole business, with role-specific streams for leadership, client services, strategy and creative, each built around briefs relevant to their client work. The leadership stream produced a documented AI strategy, the agency's first AI policy, and an AI task force with a plan and protected time to deliver it. Meanwhile everyone in the agency went through one of Spark's skill streams, learning how to apply AI to their real work, to speed up the ordinary tasks and elevate their thinking on the important ones.
What changed
AI fluency has dramatically improved across the agency, with the proportion rating themselves as highly capable with AI has gone from 24% to 75%. Meanwhile three quarters of the team report recovering three to five hours a week through productivity improvement: 10% of their working time. Proposals and awards entries that once took a week now take days.
Measured against the Spark AI Maturity Model™ the business has moved from a score of 1.5 to 2.5 with large gains across all 4 four pillars: vision and strategy, skills and culture, data and tools, governance and accountability. The changes show up across all three parts of Spark's AI strategy canvas:
Automation. Three core agents are in use with around twenty-five more in the backlog, and SharePoint and OneDrive have been re-architected so AI can Access a single version of the truth.
Augmentation. AI now supports the delivery of top-tier strategy, and Is even helping with business development. Where AI-drafted proposals had once been obvious and awkward, they now read consistently and well. And to embed the practice, AI Hour is a weekly fixture, and AI proficiency has entered job descriptions and appraisals.
Innovation. And AI is already changing the way that it delivers work to clients. For a major packaging client, the creative team is using Figma Weave to generate in situ product shots that would otherwise mean an extensive shoot repeated through the year. That work has opened a framework conversation, and it is being priced on the value delivered rather than the hours taken. On another engagement, sections of a new corporate website were delivered using Claude, with the front-end team describing their role shifting from creator to overseer.
"We built on the knowledge we had gained and started to work out where the business opportunities were. We went to a client with a different way of thinking about an opportunity we knew was there, and suddenly we were having conversations we had not been having before."
– Matthew Shannon, Executive Creative Director, SampsonMay
The return on investment
A documented AI strategy, a named task force owner, a re-architected knowledge base and AI proficiency written into hiring and appraisals
Proposals down from a week to three days, RFPs in roughly half the time, and three in four of the team recovering three to five hours a week.
A new service line delivered with AI, replacing photography that would otherwise mean a multi-day shoot repeated through the year, and a framework conversation opened off the back of it.
Stage 1 Experimentation to Stage 2 Adoption on the Spark AI maturity framework in four months, with all four maturity pillars improving.
