LivByte by Liventus Inc

The episode offers some practical framing (start with the business problem, not the tool; AI tech debt is real; the bottleneck has shifted from infrastructure to governance and process adaptation), but most of the runtime is spent on broad, well-worn AI adoption talking points. Novel ideas are lightly touched and rarely developed with enough depth to give a B2B operator actionable direction.

“the bottleneck was probably different... Now I think it's just the ability for people to adapt their processes and their governance and their learning so they, the tech can do the work”

“AI tech debt is very real. Flooding your systems and just the world, your zone, whatever you want to call it, with mediocre AI output, um, is going to hurt you more than it's going to help you”

Nearly every argument recycled from standard AI discourse: 'don't chase the shiny object,' 'start small and iterate,' 'AI is an efficiency multiplier not a replacement.' The observation that AI trained on today's data won't know tomorrow's tech is mildly interesting but underdeveloped; no genuinely contrarian or first-principles thinking emerges.

“don't chase the trend, chase the problem”

“people get so consumed with whether or not they can't. Whether or not they could doesn't necessarily mean they should”

Dan Levin is a genuine practitioner - president of a technology company, 20-plus years of enterprise implementations, actively building AI products (SparkFlow, DLN, MCP servers) - but this is essentially an internal company promotional podcast and he hasn't demonstrably deployed AI at notable external scale or produced publicly verifiable outcomes that would distinguish him from a capable mid-market technology executive.

“we're building our own MCP servers and we're building our own, you know, models and you know, for some of our products that we're developing”

“we use AI coding assistance like GitHub, uh, Copilot and Cursor and many other tools that will accelerate our development work”

The episode names real tools throughout and surfaces one concrete vertical use case (tile company with 10,000 SKUs, equipment finance touching credit bureaus, bank statements, tax returns) and a vague timing claim ('15, 20 minutes... going to be seconds'), but there are zero hard ROI numbers, no customer data, no benchmarks, and no named client outcomes - evidence stays at the anecdote level.

“their salespeople have to deal with like 10,000 different types of tile... the really good salespeople can get to it in five, six minutes. And the newer ones might, could take them an hour”

“Every lease application touches a ton of data points, credit bureaus and bank statements and tax returns and background checks and, and you know, resale values and equipment lookups”

The host interviews her own company's president, complimenting him extensively in the intro and throughout, regularly agreeing and amplifying his answers rather than probing, and never challenging a single claim. Questions are broad and often answered partly by the host herself before the guest responds, producing a promotional feel rather than a substantive dialogue.

“I couldn't agree more”

“As I mentioned in your intro, you're definitely someone, I think more so than most people I know just trying to keep up with the latest AI tools”