AI-First Organization

Mike's Strategic Framework for Applying AI in an Organization

01

Build the Infrastructure

Intelligent Data Layer

  • •Agent-traversable databases & knowledge graphs
  • •Single source of truth for team data and processes
  • •Enable stakeholder interactions and collaborations

Application Ecosystem

  • •Orchestration Agent providing users the right tools and processes
  • •Subagents for coding and knowledge synthesis maintenance
  • •Small team-specific app/agent deployments
02

Collaboration

Departmental AI Adoption

  • •Build working groups, run hackathons, partner with L&D on content
  • •Provide clear guides on prototyping and submitting ideas/needs
  • •Enable efficient SME contribution, unlocking tacit knowledge

Documentation & Scaling

  • •Subagents documenting and maintaining app/Agent functionality and state
  • •Prototype to Scale funnel
  • •Code review process enabling team-specific versions and contributions
03

Measurement

Agent Telemetry

  • •Incorporate analytics & tracing into Agents and apps
  • •Efficient SME input on traces with data pipeline creation
  • •Track usage analytics, trends, productivity, and reporting metrics

Continuous Improvement

  • •Evaluation Set creation and maintenance to improve Agents
  • •External collaboration inside and outside of AstraZeneca
  • •Benchmark solutions: determine when external agents/tools can compliment or replace internal ones

Outcomes

🚀

Accelerated Innovation

Infrastructure making prototyping to deployment fun and easy

📊

Measurable Impact

30% efficiency gain: From 10 FTEs to 7 FTEs, plus algorithm forecast improvements

💡

Novel Strategies

Pioneering Claude Code Agent applications for Knowledge Synthesis

🔄

Antifragile Development

Continual evolution that builds upon learnings and improves over time

Unlocking organizational potential where domain experts directly contribute their knowledge into AI systems