Fintech Insider Podcast by 11:FS

The episode contains a few genuinely sharp observations - particularly Sam's point on AI governance and model deprecation timelines vs regulatory requirements, and his critique of AI-assisted COBOL migration - but these are surrounded by substantial padding: generic statements about legacy systems, platitudes about speed and agility, and promotional talking points about Engine's proposition.

“the general availability models of Gemini and that kind of thing can be deprecated in weeks. If you can't process full deprecation, replacement change cycles in weeks and you can't govern uncertain outcomes in the context of all your regulatory obligations of consumer duty, vulnerable customers, that kind of thing. You won't be able to make use of these tools”

“50 different disjointed systems being migrated by 50 agents is going to produce a lot of disjointed systems that now no human understands. Rather than a few humans close to retirement understanding”

The AI-governance-and-deprecation-cycles argument and the critique of AI-led core migration are genuinely contrarian and not commonly articulated; however, the bulk of the episode recycles standard banking modernisation arguments - legacy cores bad, cloud-native good, speed matters - that have been circulating in fintech media for a decade.

“The one place I think AI is not going to live up to its potential is accelerating that tech modernization”

“50 different disjointed systems being migrated by 50 agents is going to produce a lot of disjointed systems that now no human understands”

Both guests are genuine practitioners from a credible success story - Sam has 10 years at Starling and built the tech from startup to profitable bank, Jodi leads a real commercial expansion - but the episode functions largely as a product marketing session for Engine, which blunts the practitioner value considerably.

“I've been there for 10 years now, joined a 20 something person startup”

“we were one of the first digital banks anywhere to reach profitability”

There are concrete numbers scattered throughout - the 8,000 hours/month figure, 9,000 US banks and credit unions, Chime's 9 million customers, the 60% vs 30% efficiency ratio comparison - but many claims about Engine's own capabilities, the deposit flywheel model, and the three-year transformation window are presented without supporting data or named client examples.

“Chime now has over 9 million customers”

“most mid tier banks have a efficiency ratio of around, let's say 60% or so”

The host's questions are topically appropriate and occasionally surface interesting territory, but he consistently validates every answer with phrases like 'I couldn't agree more' and 'perfect summary,' never pushes back on unsubstantiated claims (e.g. the deposit flywheel model, the three-year window assertion), and the overall dynamic is a soft promotional interview rather than a probing conversation.

“I couldn't agree more”

“I mean, it's a perfect summary”