Why I stopped waiting for AI to arrive — and started building it instead.
Most CFOs I know are waiting. Waiting for vendors to add AI to their ERP. Waiting for a consultant to tell them what to do. Waiting for the use case to become "obvious enough." I did that for a while. Then I stopped.
Over three decades, I managed finances for companies ranging from 50 to 500 people. Cash flow forecasting. Month-end close. Budget variance that nobody wanted to explain to the board. Reconciliation work that could eat a week if the data was messy. I know what the job actually looks like from the inside — not the org chart version, but the 11pm version.
It wasn't one moment. It was the accumulation of them. Every time I watched a capable analyst spend three days on a report that answered a question nobody asked the right way. Every time we built a forecast model in Excel that was already outdated by the time we presented it. Every time I realized that a huge portion of my job was process management, not judgment.
"A lot of what I did as a CFO wasn't thinking. It was orchestrating information from one place to another. AI can do that. I'd rather spend my time on the part it can't."
When I started experimenting with AI agents — systems that can pull data, flag anomalies, draft commentary, run scenario models — I wasn't surprised that they worked. I was surprised by how much of my former job description they could already cover. Not everything. But more than the industry is admitting.
Not because I have all the answers. Because I'm documenting what I'm learning in real time — as a builder, not a bystander. There's a lot of content about AI in finance that comes from people who have never closed a month, never sat across a CEO who wanted to know why marketing spend was up 18% with nothing to show for it, never had to explain a reforecast to a board with 40 minutes' notice.
This series is for finance professionals who want signal, not hype. Each post will be specific: a real use case, a real limitation, a real lesson. Some posts will be technical. Some will be about the human side — what changes when the agent does the reconciliation and you're left with only the judgment calls.
It's not a vendor pitch. It's not a "10 ways AI will transform finance" listicle. It's not going to tell you that your entire finance team will be replaced next quarter — that's not what's happening, and anyone saying so hasn't worked in a finance department recently.
What is happening: the bar for what a small finance team can produce is moving fast. Companies with 2 people in finance are starting to operate like companies that used to need 8. That gap is going to matter — for hiring, for positioning, for how CFOs justify their own role to leadership.
"The CFOs who are watching AI from the sidelines are making a strategic bet that nothing important will happen before they're ready to act. That bet is getting riskier every month."
Post #2 covers the first AI agent I actually deployed — what it does, what broke, and what I'd do differently. It's not a success story. It's an honest account of what real implementation looks like when you're not a tech company and you don't have an engineering team.
If you've been in finance for any length of time, you've seen technology promises come and go. ERP implementations that took 18 months and delivered half of what was promised. BI tools that solved last year's problem. I'm not asking you to trust the hype. I'm asking you to look at the actual work — and decide for yourself.
50 posts. One CFO's honest account of building AI into finance — from the first agent to the lessons that didn't make the demo.
Are you watching AI from the sidelines — or already experimenting?
No spam. One post per week. Unsubscribe any time.