Fireside with Founders & Leaders

There are a handful of genuinely useful practitioner observations - pilot purgatory, the baseline requirement for proving AI ROI, the need for daily token usage monitoring - but they are buried under extended filler about cultural immersion in Austria, repeated exhortations about 'process and structure,' and generic leadership platitudes that consume the majority of the runtime.

“some have gone into what I would refer as pilot purgatory right where they've got 300 pilots of whatever it is that's running in the background burning a huge amount of tokens”

“if you're gonna restructure your way of working you actually have to have a baseline cause otherwise you're investing in AI but you're if you don't if you don't know where you're at as a baseline how can you ever prove the value of the investment”

The episode recycles the standard 2024 CTO conference playbook - tech must align to business outcomes, generative AI is tactical while agentic is strategic, cloud adoption analogy - with only occasional fresh framing such as 'pilot purgatory' or the observation that AI consumer familiarity reduces internal resistance to agentic pitches.

“the generative stuff is very reactive very tactical it's not gonna really be the thing which moves the needle in your business the agentic stuff potentially is”

“we used to have cloud first now we've got AI first”

David Burns is a legitimate practitioner who has run large-scale technology transformations at HP and CGI and is now CTO of a real PE-backed business implementing agentic AI; however, TLC Worldwide is a mid-market loyalty-marketing company and he has not built or exited anything of notable scale, placing him solidly in the competent-operator tier rather than the exceptional one.

“I've got that kind of grounding in real good solid processes and ways of working”

“a lot of recruitments of Ctos and also chief product officers and various other uh key staff is done post a deal yep because sometimes there is a need for a step change sometimes the CTO that got you to here is not the CTO which is gonna get you to there”

The episode name-drops Anthropic, OpenAI, Langchain, MCP, and RAG and references a recommendation engine with roughly 90% accuracy, but almost no verified metrics, client names, dollar figures, or before/after comparisons are provided; most examples remain illustrative and vague.

“around about 90% 8% accuracy”

“they're literally talking about a valuation which is just short of $1 trillion”

The host occasionally lands a useful follow-up - pushing on the due-diligence versus post-deal distinction, asking whether a human could match AI prediction accuracy - but largely validates every claim, lets tangents about lederhosen and Austria run long, and never challenges a single assertion, resulting in a PR-adjacent chat rather than a probing interview.

“that's so key that is so key”

“would you categorize those things under due diligence or would you separate those two things”