Still in the seat.

When I was a CFO, I:

• Raised tens of millions of dollars
• Opened 30+ health practices to provide care for skilled nursing facilities across the country
• Helped hire 40+ people
• Ran diligence to buy several companies
• Audited financials and found executives stealing money
• Coached CEOs, COOs, and CROs on how to build scalable systems

Can AI replace a lot of the tedious work that falls under the CFO, like FP&A? Absolutely. But if you're a real CFO doing real CFO stuff, you're not directly doing that work anyways.

The new "AI CFO" companies will disappear as quickly as the dozens before them that claimed the same.

Selling has never been my exit plan, but this was a huge company that had tried to reverse engineer what we built for years and realized they couldn't. I was told the acquisition was their highest priority.

We went through every piece of diligence: business, market, legal, tech, and team. For months. Then a new corp dev middle manager got added to the mix, and the friendly conversations about terms stopped being friendly.

After weeks of negotiating with an uneasy feeling, I walked away and kept building independently. A few weeks later that entire department blew up.

If the deal had gone through, I may have been left without a company and without a job. It was a disappointment at the time. It turned out to be a blessing.

It takes no time, and Claude can do it for your company or your entire client base. If all you need are charts, graphs, and analysis of what already happened, there's no reason to have an analyst or a piece of software doing it for you.

Reporting has always been a check-the-box feature. The real value for a company is everything forward looking.

If a prospect comes to me asking for reporting and reporting only, I tell them to use Claude.

A founder called me the week before payroll. The gap was a signed contract they were treating as cash, for work they hadn't delivered yet. It wasn't cash to spend.

We spent two days rebuilding the forecast off collections instead of bookings. He made payroll, then called off two hires he was about to sign.

Nobody warns founders that revenue and cash are two different animals. They find out the week before payroll.

You ask a question, it runs a canned query, and it hands back a number you could've pulled from a pivot table yourself. Calling that AI is generous.

Claude gets closer than most finance software I've seen, and it wasn't even built for finance. That should worry everyone in this space, me included.

That's why our focus at Clockwork is being un-Claudeable.

The client paid for the day. That was every single month when I ran my outsourced CFO practice.

For two years I was convinced it was a pricing problem. I raised prices twice and rewrote the proposal more times than I want to admit. It was never the price. It was the week.

If you're adding advisory and the math isn't working, count how many hours you eat before anyone actually gets advice.

For two years I priced advisory from my cost up. Hours times rate plus a margin. It always came out too low, because the client was never comparing me to my hours. They were comparing me to what happens if they get the decision wrong.

A $4M business about to hire 3 people and sign a lease has $600K of decisions on the table this year. $4,000 a month to get those right isn't a hard conversation. $4,000 a month for "monthly reporting" is.

The pattern I've seen for over 10 years: a consultant who never owned a P&L. Someone who never sat in the seat they're selling, with no payroll they had to make, no covenant to renegotiate, and no board to face. Strictly backward looking: last month, formatted nicely and called strategy.

Meanwhile operators are making life or death calls for the company. Hire or wait. Raise or bootstrap. Take the offer or keep building.

A real CFO doesn't hand you last month's numbers. They tell you whether the light at the end of the tunnel is daylight, or a train.

The worst mistake I see firms make is hiring an analyst to launch advisory services. Nine times out of ten that person doesn't work out, and the firm burns nine months and tens of thousands of dollars.

In the age of AI, the skill gap between a bookkeeper and a full-blown advisory executive is much smaller.

Launch it with the team you have and figure it out as you go. You already know what to do.