The Founder's Formula Podcast

The episode has a few genuinely useful observations - LLMs used as interviewers rather than data-query tools, the distortion of customer voice as companies scale, and the gap between CRM disposition data and reality - but these are buried under extended biographical backstory and the host's own lengthy tangential anecdotes. Novel ideas per minute is low.

“I wanted to see if I could actually write a prompt and engineer a couple different bots simultaneously to play the role of an interviewer, a user interviewer”

“tools and employees got between them and that end customer and ultimately the voice of the customer got filtered and distorted”

The framing of an LLM conducting structured user interviews (rather than answering questions) has some freshness, and the observation that cross-functional leaders can't argue against direct customer voice is a decent insight. Everything else - win-loss analysis, sales reps blaming product/budget, co-founder dynamics - is well-worn territory.

“can I leverage the LLM to extract new insights from humans rather than asking it to go extract insights from a corpus of data or the Internet”

“who can't we argue against? Well we can't argue against the customer. So we get direct raw feedback from them in aggregate”

Brandon Santos is a credible first-time founder with relevant sales-leadership experience at mid-market SaaS companies, but has not led a go-to-market function at meaningful scale and is at a pre-revenue MVP stage. As a portfolio company of the sponsoring VC, the episode has a promotional character that limits candor.

“I actually recruited him into Center Code to be my sales manager”

“we're very early on in this thing is finding design partners. Not even trying to get customers, necessarily”

A handful of named companies and resources (Blue Jeans, Center Code, Mavenlink, FSK Ventures, YC co-founder matching, CoFounders Lab) and one close-rate figure (22% to 35%, from the host not the guest) add texture, but the product has no live customers or performance data, and most claims about impact are hypothetical or aspirational.

“close rates go from 22 to like 35”

“We've been working with them for the past five months”

The host repeatedly derails into multi-minute personal anecdotes (3PL Central win-loss program, technical co-founder search woes) that crowd out the guest and go unconnected to sharper follow-up questions. Questions are mostly leading and softball, with no meaningful pushback on product claims or go-to-market assumptions.

“we'd pay a pretty hefty penny to a company to manually call, um, closed lost deals, figure out why they didn't buy and we go through the whole report”

“I got totally hosed on something I was trying to put together and I had somebody invoicing me for work that wasn't even done”