GrowCFO Show

The episode delivers concrete operational insights about AI deployment failures and successes, particularly around task automation vs. full workflow automation, the three deployment models (augmentation, task automation, application development), and specific ROI metrics. However, it contains significant padding including lengthy introductions, promotional content for Grow CFO, and repetitive validation between hosts that dilutes the substance-to-filler ratio.

“Task automation is really a thing and we've been able to leverage that brilliantly here at Envoy, something that was taking us about two hours to do on a monthly basis. I took it down to two minutes”

“Full workflow automation is not really a thing right now in most of the work that we do, ah, at finance. But task automation is really a thing”

While the three-tier deployment framework (augmentation, task automation, application development) is a useful organizational model, it is not particularly novel or contrarian. The core arguments about IT gatekeeping, the need for CFO technical literacy, and MCPs enabling customization are reasonable but largely represent conventional wisdom emerging in the AI-for-business discourse. The guest recycles common themes about hallucination bugs and API limitations without fresh theoretical insight.

“One is augmentation, so you can Augment things like I'm sure you augmented yourself for this interview by using ChatGPT or Claude to do something. The second one is task automation...The last one is what I call application development”

“I see this as a today problem. I think that the future, quite frankly is going to be in MCPS and APIs being the broad distribution of information for CFOs”

Sinohe Terrero is a highly credible guest as CFO and COO of a Series C company (6,000 customers, $250 headcount, significant ARR), with 20 years in tech and prior experience as a serial startup CFO. He demonstrates genuine hands-on practitioner experience building AI applications, leading AI governance councils, and shipping measurable ROI. This is substantially more credible than typical podcast guests who discuss AI theoretically.

“I'm now what I consider a serial sort of startup cfo. Been doing this for a long time”

“We are a series C company...have about 6,000 customers across the globe...about 250 folks, uh, most of us here in San Francisco”

The episode provides concrete examples (bank rec automation from 2 hours to 2 minutes, $40,000 identified in insurance reconciliation, displaced tools like Floqast/Asana/Monday.com, Morning Coffee dashboard, payroll health insurance PDF reconciliation) and specific tools (Claude, Scribe, Granola, MCP). However, many claims lack supporting metrics: no revenue impact quantified, no timeline specificity on tool displacement, no customer impact data, and vague references to "hundreds of thousands of dollars" saved without breakdown.

“something that was taking us about two hours to do on a monthly basis. I took it down to two minutes”

“She built a skill where now she's feeding both sides of the equation, feeding the payroll information in detail deductions and feeding the PDFs from the companies...we were able to identify $40,000 that we were owed”

Kevin Appleby asks reasonable follow-up questions and pushes back on a few points (API hallucination issues, headcount reduction), but the conversation lacks sharp challenge or productive disagreement. He often validates the guest's points rather than interrogate them, misses opportunities to probe on governance tradeoffs, doesn't ask for harder metrics on ROI, and allows vague claims ("hundreds of thousands") to pass unchallenged. The exchange is collegial and mutually reinforcing but not incisive.

“Does the IT department get in the way when you're doing this?”

“When you're asking it to read emails and delve into the spreadsheets that are attached to the emails, is it generally coming back with the right answer?”