Transformation Ground Control

The episode has a handful of genuinely useful reframes - ERP as system of record vs. AI as system of action, AI changing the order of investment - but these are diluted by the host spending enormous portions of the runtime restating what the guest just said and delivering extended solo editorials. The insight-to-runtime ratio is poor for a 97-minute episode.

“manufacturers don't have an ERP problem. Nobody's really talking to me about that. They've got a fundamental ERP problem. I think what they really have is a decision making problem.”

“maybe 20 years ago, ERP replaced paper, today AI is replacing clicks. So literally it's kind of sort of the way of engagement is changing.”

The 'service as a software' inversion is a clever phrase, and framing manufacturers as having a decision problem rather than an ERP problem is a genuine reframe. However, the surrounding arguments - vendor incentives misalign with customer interests, one-size-fits-all cloud is oversold, on-prem isn't dead - are well-worn independent-consultant talking points with little novel first-principles development.

“we're moving to a world which is more service as a software rather than software as a service”

“AI has really changed the order of investment. If you think five years ago companies were thinking about, hey, first we'll invest in infrastructure, then we'll invest in applications and analytics, and then maybe somewhere we might do something with AI.”

Sanjay has genuine operator depth - Honda shop floor trainee, 16 years at IBM in manufacturing, CEO of billion-dollar Software AG through a turnaround, now CEO of QAD. The credentials are real and directly relevant. The discount is that he is transparently selling his own product throughout and the host never creates conditions that would stress-test his claims.

“Prior to that I was the CEO of another software company called Software AG is a German multinational company doing database and integration application, uh, products, roughly about a billion in size, uh, 5,000 people”

“I started my career in Honda, you know, as an, uh, as a trainee working um, on the shop floor assembling engines.”

There are concrete data points - 50% to 93% software usage at a named customer (Tenacle), 15-25% inventory carrying cost reduction, 40-50% buyer time savings, 8 agents with specific SKUs, 90-day deployment methodology, 7-10% of opex IT budgets carved for AI from a 15-exec roundtable. These are better than pure abstraction but are all self-reported by a vendor with obvious commercial interest and rarely independently corroborated.

“when we look at the deploys and how they've gone, uh, the engagement or usage of the software has gone from 50 to 93%”

“customers can actually reduce inventory carrying costs by 15 to 25%”

The host asks a few pointed questions (on SAP lock-in, on commoditization of ERP by AI layers) but habitually answers them himself before the guest can, then pauses the interview to deliver extended solo editorials repeating the guest's point. There is zero pushback, no challenge to any claim, and the dynamic is clearly promotional - host visited guest's office and had dinner with him the night before recording.

“Yeah, yeah, very interesting. Um, I'm curious to hear from the audience too.”

“I want to stop for a second here because Sanjay just had a really good point. It's very contrarian but very true.”