Two Cents: Finance Talk

The episode covers useful foundational principles (product mindset, starting small, foundations matter, actionable vs. vanity metrics) but relies heavily on reiterating the same 3-4 core ideas throughout. While substantive, there's limited novel insight per minute - much time spent on context-setting and gentle agreement rather than new frameworks or counterintuitive observations that a finance operator wouldn't already suspect.

“Don't go for vanity metrics. Don't go for some of the flashy metrics, but go for some really actionable metrics.”

“AI is always a living capability.”

The core thesis - apply product thinking to AI, fix foundations first, iterate with users - is solid but well-trodden advice in the AI/product world. The specific framing for finance is useful, but the frameworks (MVP, product mindset, human-in-the-loop) are standard. The guest does not present contrarian views, first-principles reasoning, or novel mental models that challenge conventional thinking.

“building AI the product way, usually think of it as a switch in mindset”

“AI is always a living capability”

Strong caliber: Abhishek Chandna is a Director of Finance at Visa with prior experience leading transformation at Autodesk - genuine operator at scale in the domain. He speaks from hands-on experience running AI initiatives and building copilots for actual finance teams. This is not a consultant or academic; it's a practitioner with real accountability and lived problems.

“I'm currently working as a Director of Finance, Data Strategy and Governance at Visa”

“I'm doing within my team”

The episode includes some concrete examples (variance analysis commentary, comparing Excel files, Copilot usage, Satya Nadella quote) but lacks quantified outcomes, metrics, or timelines. No data on success rates, adoption rates, cost savings, or specific dollar figures. The variance analysis example is well-explained but remains illustrative rather than evidenced with measurable results or comparable case studies.

“variance analysis commentary every month, every quarter”

“we have to compare to Excel files. That's a pretty regular work that we do”

The host (Anthony) asks reasonable setup questions and does attempt some follow-up (e.g., on legacy systems, on ethics/human oversight balance, on AI use cases). However, questions are mostly softball and rarely challenge the guest's claims. The guest is rarely pushed for specifics, counterexamples, or failure modes. The conversation feels collaborative but lacks the productive friction or sharp follow-ups that would elevate it.

“Yeah, I like that a lot. It's kind of moving away from that mindset”

“Yeah, definitely. Yeah, it seems like so. Or, or I think that's good advice as well”