This episode features CIO Brett Johnstone sharing Pitcher Partners’ journey implementing Microsoft Copilot and custom AI solutions.
Brett’s AI Implementation Strategy
Brett categorises AI at Pitcher Partners as:
- AI in line-of-business apps
- Microsoft Copilot for productivity
- Custom apps/agents (their main focus)
About half the firm uses Copilot, with 30 new users onboarded monthly through structured workshops. Brett emphasises Copilot excels at just three functions (search, summarisation, content generation) and shifted from seeking “killer use cases” to finding “1 million $1 ideas” – small time-savings that justify the monthly cost.
For custom AI, they’re building proof-of-concept applications using RAG models, focusing on narrow-domain agents to reduce hallucination risk. Brett advocates that “the narrower the subject matter, the more accurate” the AI.
Implementation Advice
Brett warns against paralysis from data governance concerns, suggesting controlled implementation with appropriate policies. He believes AI is “horribly overhyped” with expectations that are “too high and timelines too short,” but sees its impact on the CIO role as evolutionary rather than revolutionary.
The key to success has been their investment in “prompt training” with a custom framework, regular learning sessions, and encouraging users to share successful use cases to create network effects.
