Framework · free to use

How to pitch AI to your board.

If you are a CEO or CFO at a mid-market B2B company and you have to take an AI investment thesis to the board, the rest of this page is for you. The honest version, not the hype version — what to include, what to leave out, and how to frame the ask.

Frame the ask

Productivity, not transformation

"Transformation" makes boards nervous. "Productivity improvement" makes boards nod. Frame the AI investment as a measurable productivity initiative with defined ROI — not as a strategic transformation requiring sustained capital.

Concrete framing: "We're investing \$X this year in AI tooling and a structured rollout. We expect Y measurable productivity gains in 6-12 months across these specific workflows."

The 7-slide deck

What to actually present

The opportunity — one slide. "AI productivity gains at mid-market companies in our segment are running 15-30% in functions where it's been deployed. We have not captured this yet."
What we will do — one slide. "Roll out AI across 3 workflows in sales/marketing/ops over 90 days, then expand."
The numbers — one slide. Cost (license, implementation, internal time) vs. expected benefit (hours recovered, throughput gain, avoided hires). Realistic, not hype.
The risks — one slide. Data handling, brand voice, accuracy. How each is mitigated.
The team — one slide. Who owns it. How much time. Why this person.
What we need from the board — one slide. Specific ask. Budget approval, policy approval, etc.
The 90-day milestones — one slide. What success looks like at 30, 60, 90 days, so the board can hold you accountable.

What boards actually ask

Be ready for these

"Why now?" Because competitors are 12-18 months into doing this and we are not. The cost of waiting compounds.
"What if it doesn't work?" The downside is \$X in spend and a quarter of attention. Bounded. The upside is 5-10x that in recurring productivity.
"Who's accountable?" Name the person. Show their experience or external support.
"Are we exposed to data risk?" Walk through the policy and the tier you're using. Specifically address training-on-input concerns.
"Should we hire someone for this?" Usually no for the rollout phase. Yes for sustained operations at \$30M+.

What to leave out

Cut from the deck

AGI / Skynet / existential framing. Distracting at the board level.
Hypothetical revenue projections. Boards distrust them.
Vendor logo soup. They do not care which 12 vendors you considered.
Anything about "AI maturity model" frameworks. Generic, not credible.
Photos of robots. Pretty please.

Realistic numbers to anchor to

What the board will accept

Year 1 spend at \$5M-\$30M company: typically \$15K-\$80K all-in (platform + implementation + training).
Productivity gains: 15-30% in the functions where AI is deployed, in months 6-12.
Time to ROI: typically 6-9 months for measurable hours-saved ROI; 12-18 months for compounding capability ROI.
Headcount implications: usually one avoided hire per \$5M of revenue scale within 12-18 months.