Episode Summary

In this episode of TechUnhinged, host Ashar Samdani sits down with Paul Lewis, CTO at Pythian, a data, cloud, and AI services firm. Paul argues that AI advantage does not come from buying tools; it comes from the operating model built around them.

They cover enterprise AI adoption, why replaceability now outranks availability, why discrete agents close to the source beat one super agent, and governance before the first deployment. Paul also explains why asking business teams to build their own no-code agents backfires and creates a succession problem.

Key Takeaways:

  • Why enterprise AI adoption stalls on the operating model, not the tool
  • Why replaceability is now the top requirement for AI in production
  • What happens when business teams build their own no-code AI agents
  • Why discrete agents close to the source beat one enterprise super agent
  • How to separate real AI use cases from analytics and deterministic workflows
  • Why AI governance has to be in place before the first deployment

Guest Bio:

Paul Lewis is Chief Technology Officer at Pythian, where he leads technology strategy for data, cloud, and AI programs. Before joining Pythian in 2021, he was Global CTO at Hitachi Vantara, after an earlier career in financial services that included nearly a decade of executive roles at Davis + Henderson. With more than three decades in IT, he is a CIO and CTO analyst at GigaOm and serves on advisory boards for Schulich School of Business’s Master of Management in Artificial Intelligence and Master of Business Analytics programs.

Resources: