Alex Lieberman@businessbarista
I spoke to the Head of Finance for a Series B tech startup yesterday.
He did a full AI show-and-tell & by the end it felt like he was going to bust through the screen like the Kool-Aid man.
Some highlights from the call:
1. The first problem he attacked with AI was building a financial operating model for investors to raise their next round of funding.
His process: Uploaded data room context from previous fundraise, had Claude ask clarifying questions, then iteratively built ARR waterfall and expense budget.
He said he finished the model in 2 hours vs 1-2 weeks (without AI), and when I asked him about hallucination, his response was “once, and I don’t think it was the models fault.”
2. The next problem he tackled was on the customer success side.
This company sells into large institutions, and “what’s the ROI” is constantly discussed on customer calls.
He built a dashboard for clients that tracks incremental revenue (reported by customers), contract cost, and other important ROI metrics.
He also made an internal version of this dashboard that notifies customer success when ROI dips below a certain threshold.
3. He took the ROI dashboard one step further and added a function that exports key metrics and charts from a customers dashboard into a brand-aligned QBR deck that is comprehensive and editable.
4. He brought up this concept of the AI stack I can’t stop thinking about.
Traditionally, when you hire an executive, you don’t just hire them for their qualifications. You also hire them for their network, playbooks, and vendor preferences.
The same thing will happen for ALL employees. When you hire someone in the future, you’ll hire them for their AI stack (their internal tools, workflows, and skills), which gives them leverage and edge.
5. He believes the painstaking task of month end close for a company’s financials can be taken from 10 days to 1 day with the right AI systems.
6. Every knowledge worker is turning into an engineer, and the way every knowledge worker approaches their work will look like the software lifecycle.
This guy talked about his approach to building financial models with Claude, and it sounded exactly how my engineers build software.
Context dump -> PRD -> master plan -> orchestrate/execute -> iterate
7. The best way to rebuild your company processes to be AI-native is NOT to ask employees for problems. It’s audit everything they do.
Sit behind them as they work. Take a fine tooth comb to their calendar. Have them walk you through their most common workflows.
Play offense not defense in AI-transformation.
8. Final project: he built a compensation management system by using Google Sheets app script macros generated by Claude Code.
Its capabilities include system pulls from Rippling API and financial model for scenario planning, auto-generated presentation slides with compensation philosophy and individual comp details, createing permissioned folders for each manager with team-specific compensation documentation, and leveling band validation (flags out-of-band compensation.
P.S. I spend half of my week talking to execs about how they use AI in their business.
If you’re an exec and you’re afraid your company isn’t doing enough with AI, shoot me an email and my team will help create a list of clear AI opportunities specific to your business.
alex@tenex[.]co
