I’ve sat in on many budget meetings where a capital plan looked fantastic on the screen. Clean charts. Every asset scored. Twenty years of replacements, sorted by priority. Then a board member asks one simple question: “Why is the gym roof ahead of the boiler at the middle school?” And the room goes quiet, because nobody can explain it. More and more, the honest answer is “the AI ranked it that way.”
That plan is dead the minute the question lands.
AI is a great tool for capital planning. I use it, and you should too. But here’s the thing nobody selling it will tell you: AI can make a plan faster. It can’t make a plan defensible. Only you can do that.
What “defensible” actually means
A defensible capital plan isn’t the one with the best software behind it. It’s the one where you can answer three questions about any line, on the spot:
- Where did this number come from? An inspection, a work order history, a vendor quote, an engineer’s estimate. Something real.
- Why is it ranked here? Because of the consequence if it fails, scored the same way you scored everything else.
- What happens if we don’t fund it? In plain language: which building closes, which service stops, what it costs.
If AI helps you answer those faster, great. If it gives you answers you can’t trace, it just made your plan easier to attack.
Where AI earns its keep
Most of the time in capital planning goes to grunt work, not judgment. That’s where AI shines:
- Cleaning up the asset register. “AHU-1,” “Air Handler #1” and “AHU 01 Gym” are the same unit. AI is very good at finding duplicates, standardizing names and flagging missing install dates and sizes.
- Reading the stuff nobody has time to read. Ten years of condition assessments, roof reports and inspection PDFs. AI can pull the deficiencies, the ratings and the recommended actions into one table in an afternoon.
- Mining the work order history. Which assets eat the most repair dollars? Which ones keep failing the same way? That’s your evidence for repair versus replace.
- Running the scenarios. What does the backlog look like in ten years at current funding? At 2% of replacement value? With a bond? AI can build and explain those models fast.
- Drafting the story. Turning a spreadsheet into a one-page summary a board member can read in five minutes.
Use AI to do the work faster. Never use it to make the decision for you.
Where it will burn you
Watch out for these:
- Made-up numbers. Ask an AI tool for a replacement cost and it will give you one, with total confidence, whether it has a source or not. If you can’t point to where the number came from, it doesn’t go in the plan.
- Black-box scores. A risk score of 87 means nothing to a board if you can’t explain how it was built. Your criticality and consequence framework has to be set by people, written down, and applied the same way to everything.
- Garbage in, polished garbage out. AI doesn’t fix a bad register. It makes a bad register look professional, which is worse.
- Skipping the field. No model knows the chiller was rebuilt in 2019 or that the “spare” breaker feeds the science lab. Dave does.
A workflow that holds up
Here’s the approach I use and teach. AI does the heavy lifting in every step. People own every decision.
- 1. Set the rules first. Decide how you score condition, consequence and risk before you touch the data. Write it down. This is your framework, not the AI’s.
- 2. Let AI clean and assemble. Register cleanup, document extraction, work order analysis. Keep a source for every data point.
- 3. Do the Dave check. Pick 20 assets your senior people know cold. If the AI’s picture doesn’t match what they know, find out why before you go any further.
- 4. Score and rank by people, using the rules. AI can apply the framework, but a person reviews the top of the list and signs off on it.
- 5. Keep a decision log. Every time you move something up or down by judgment, write down who and why. That log is what saves you when the board asks.
- 6. Let AI draft the story, then you tell it. The board hears it from you, in your words, with your name on it.
The bottom line
The board doesn’t care what tool you used. They care whether you know your assets and whether you can explain your choices. AI can give you back weeks of grunt work so you have more time to do exactly that.
Just don’t ever stand up in front of the board and say “the AI ranked it that way.” You better be able to say “we ranked it that way, and here’s why.”
How are you using AI in your capital planning today? And could you defend every number in your plan if the board asked tomorrow?
