Most AI rollouts start with the tools and wonder why nothing changes. In this episode, Andy Lambert, co-founder of AcademyAI, talks to Rich Merrett, Head of AI at Bellrock, about the framework he has used to AI-enable two businesses, first at Citation and now at Bellrock.

Rich's starting point is that AI is an 80% people problem and a 20% tech problem. His strategy rests on four pillars: governance built around a yes-first AI policy rather than an infosec policy wearing an AI T-shirt, a people pillar that runs company-wide training, leadership workshops, hands-on build days and self-selecting champions all at once, a solutions pillar for building what people need, and a monitoring pillar that most organisations miss.

Rich explains how a sales callback tool built at Citation made a quarter of a million pounds in its first four weeks and won the National AI Award for innovation in 2025. He walks through his week of half-hour sessions, open to the whole business every nine months, ending with an "art of the impossible" day where colleagues show what they've built. He covers the four tiers of risk in Bellrock's AI policy, the 10-day feasibility sprints his team runs, why being caught out by the GPT-4 deprecation led him to monitor for model drift and bring subject matter experts back every six months, and how a Teams chatbot feeds hundreds of ideas into a pipeline that Claude scores on impact, strategic fit, time to value and reach against effort. He also talks about matching tools to job role and skill level, calling out objections like sustainability and hallucinations before anyone raises them, a new 12-level badge system for Bellrock's 71 champions, and why active reading may be the most important skill of the next decade.

Chapters
00:00 Introduction
00:52 From teacher to Head of AI
02:21 Why AI is 80% people and 20% tech
03:34 The award-winning sales tool at Citation
05:26 Pillar one: a yes-first AI policy
06:00 Pillar two: people, starting with a week of open sessions
09:07 Why champions must be self-selecting
11:43 Pillar three: DIY wins, 10-day sprints and transformation projects
12:57 Pillar four: monitoring ROI and model drift
14:48 Minimum viable AI governance
17:06 Who should own AI?
19:58 Top down, bottom up and inside out
21:00 Calling out the elephant in the room
26:23 Making the training stick
28:25 Matching tools to role and skill level
33:07 Scoring hundreds of ideas with Claude
36:22 Keeping champions engaged, and 12 levels of badges
41:30 Why active reading is the skill of the future

Connect with Rich
LinkedIn: Richard Merrett

Learn more about AcademyAI: academyai.co

#AITransformation #AIStrategy #AIGovernance #AIChampions #FutureOfWork