Everyone talks about how AI saves time. That’s only part of it.
Two ways I use it
Execution. Drafting investor updates, the first pass of a board memo, summarizing dense documents, turning messy notes into an outline. AI gets me about 40% of the way there in a fraction of the time. It doesn’t finish the job. It compresses the low-value parts of high-value work.
Inquiry. This is where the real advantage is. I use AI as a sparring partner:
- What’s the second- or third-order effect of this decision?
- What root cause am I missing behind a metric that’s off?
- How would this look with six fewer months of runway?
- How would the board, or a key employee, read this decision?
It exposes bad reasoning
When you’re running lean, the biggest risk isn’t a bad model. It’s bad reasoning that looks clean. I regularly ask AI to pick apart a decision I’ve already made. Sometimes it confirms I’m solid. Often it shows me where I oversimplified, assumed too much or skipped a step.
That’s also why some people find AI frustrating. Vague thinking in, vague output out. AI doesn’t cause bad decisions. It reveals them. You don’t need fewer tools. You need sharper questions.
Speed is table stakes. Depth isn’t.
Everyone can draft a deck or summarize a call in seconds now. Depth is still rare: using AI to model how a hiring or pricing change ripples through runway, or to test whether you’re solving the right problem or just the obvious one.
Where I start with founders
- Board and investor updates: AI drafts from the numbers; a person checks every claim.
- Variance analysis: AI flags what moved and suggests why; a person confirms the why.
- Scenario thinking: AI stress-tests the plan against a bad quarter.
- Reconciliation and categorization: AI does the matching; a person approves.
The line I won’t cross
I automate a lot. I don’t let AI make the decision.
Should we delay hiring to extend runway? Is this pricing change worth the churn risk? AI can model the scenarios and stress-test the logic. The call, and the accountability, are still mine. There’s a difference between automation and abdication.
AI won’t replace good judgment. It will show you very quickly whether you have it.
Common questions
Is it safe to put company financials into an AI tool?
It depends on the tool and its settings. Use business accounts that don’t train on your data, limit access to the people who need it, and keep anything sensitive, like payroll detail or investor terms, out of tools you haven’t vetted.
Will AI replace my finance team?
It will change what they spend time on. Data entry, matching and first drafts shrink. Review, judgment and decisions don’t. Small teams that use AI well can do work that used to need more people.
Where should a founder start?
Pick one recurring task that eats hours every month, like the investor update or variance commentary. Have AI draft it from your real numbers, then review every line. Once that works, move to the next task.
How do I know if the AI got it wrong?
Check it against something you trust: last month’s actuals, the bank balance, a number you know by heart. If you can’t tell whether an output is right, that’s a sign you need someone with more finance experience reviewing it.
AI for Finance