Know Yourself Before You Transform: AI, Decisions and the Organization
A bank CTO on why transformations fail before they start, and what is left once AI makes execution instant
Podcast Summary
In the 8th episode of the AI x DevOps podcast, host Rohit Raveendran sits down with Sailesh Panchal, CTO of TSB, a UK bank with around five million customers. Sailesh has spent over three decades leading and observing transformations, from CTO at Travelex to helping Lloyds Banking Group on its journey to cloud. His argument: most transformations fail before they start, because the organization doesn't know itself.
Know Yourself Before You Transform
The real primitives are decisions, authority (and delegated authority), inheritance, jobs to be done, and value. Most organizations can't write these down, which is why partners end up doing the excavation work.
Every Consequential Change Is a Transformation
Organizations run many transformations at once, from HR programs to cloud moves, and the discipline underneath them has to be consistent.
Mine Your Decisions With AI
Sailesh used AI to extract hundreds of decisions from existing documents into a reviewable, structured record, then into a graph you can project future states from. He also walks through MTDR, his open-source standard for Transformation Decision Records.
DevOps Was an Organizational Redesign
Deployment velocity is meaningless if the organization isn't ready. Features hide behind flags until the company catches up.
Great Engineers Become Great AI Mentors
The skills that made someone a strong lead engineer (code review, mentoring, explaining why) are exactly the skills that make AI useful in engineering.
The Zero-Time Thought Experiment
If you could execute everything instantly, what is disaster recovery? What does change approval mean? Start with a glass-tube pilot and redesign the roles around it.
More Live Content
View allRelated Articles
View allFacets Cloud's AI Assistants for Infrastructure as Code Recognition Now Spans a Market Guide and 5 Gartner Hype Cycles
People-Led vs Facets-Led Cloud Migrations
Runtime Behavioral Nudging for Large Language Model Agents

