The Importance of Employee Education in Adopting AI to Drive Meaningful Change

Prefer audio?

Who should listen: Microsoft 365 service owners, modern workplace architects and adoption/change leads planning or troubleshooting a Copilot rollout, plus partners building AI adoption practices.

Guest: Jamie Wheeldon — Chief Architect, Symity

Jamie Wheeldon, Chief Architect at Symity and a Microsoft Copilot jumpstart partner, joins Tom Arbuthnot to place generative AI in the context of decades of earlier AI work and to explain why licence purchases alone rarely deliver value. The conversation covers Symity's 8–12 week adoption sprint model, data and IP concerns, and how the growing Copilot SKU landscape maps to real use cases.

Key insights

• Two distinct customer patterns are emerging: organisations investing properly in adoption campaigns and thought leadership, versus those buying half a dozen licences to 'have a play' — the latter typically stall because they don't know what they don't know. ▶ 2:03
• Copilot deployments that stop at meeting recap are underwhelming. GitHub Copilot has been notably more successful because developers inherently understand the value, and Copilot in Power BI removes the need to hand-write complex analytics queries. ▶ 3:36
• Generative AI is not new technology — it sits on rule-based expert systems from the 1970s, fuzzy logic, robotics AI, neural networks, machine learning and natural language processing. Framing it that way helps set realistic expectations with customers. ▶ 4:38
• Different AI platforms treat submitted data very differently: some enforce strict data security, others use inputs to train their models, raising the question of whether your data can surface in someone else's query. Understand the data handling before allowing tools in. ▶ 10:19
• Wheeldon cites Bloomberg market research forecasting AI growing to a $1.3 trillion market within 7–8 years, with AI in IT services going from roughly $83 million to $86 billion over the same period — useful context for why every vendor is racing to market. ▶ 12:21
• Previous technology rollouts were evolutionary — 'press these buttons'. Generative AI is transformational: users get a blank box, so training must cover prompt construction, spotting bias and critically reviewing output rather than button-clicking. ▶ 20:03
• Wheeldon's mental model for users: treat generative AI like an intern — give clear instructions, expect iterative back-and-forth rather than a perfect first answer, and always review what comes back. ▶ 21:35
• Symity's Copilot for M365 adoption framework runs as 8–12 week sprints per user community: education up front, drip-fed video training, weekly meetings, usage tracking fed back to users, and a Centre of Excellence for sharing tips. Without sustained engagement users revert to old habits within a week or two. ▶ 22:36
• Microsoft's Prompt Buddy app is recommended for capturing organisation-specific prompts and use cases and sharing that internal knowledge across teams. ▶ 24:08
• Copilot is a suite, not a product — Arbuthnot counted around 14 separate 'Copilot for' SKUs and roughly 19 'Copilot in' features embedded in larger SKUs, and the list keeps growing. ▶ 25:08

Insights summarised by AI from the episode transcript, reviewed by the Empowering.Cloud team.

Listen: Apple Podcasts · Spotify · Other platforms