How to pitch AI to your boss

Why does my boss keep saying no to AI?

Because the pitch is answering the wrong question. Most AI proposals argue that AI is valuable, but your boss already believes that. What they are actually weighing is risk to their own credibility: approving your project means attaching their name to its cost, its timeline, and its failure mode, and a proposal that is vague on any of those reads as career risk, not opportunity.

In our conversations with operators and consultants inside mid-size and large companies, one rule held six times out of six: an initiative dies unless it resolves to a hard top-line or bottom-line dollar figure. Enthusiasm without a number does not compete with the funded projects already on the table. The no you keep hearing is rarely a verdict on AI. It is a verdict on the packaging.

What does an executive actually need to see?

Four things. Call them the Executive Four:

• Cost. The all-in number for the pilot: software, people’s time, and any outside help. One figure, not a cloud of maybes.
• Payback period. When the benefit catches up with the spend. Most executives will wave through a small operational investment that pays back inside 12 months; past that, expect a real review.
• An owner. A named person who runs it, reports on it, and is still accountable in six months. “The team will absorb it” is how proposals die.
• The failure exit. What gets decided, and when, if it does not work. A defined stopping point turns an open-ended risk into a bounded experiment.

Bring those four and you have changed the conversation from “should we believe in AI” to “should we approve this specific, bounded spend” – a question executives say yes to every week.

What should I ask for?

A pilot, not a program. One process, one team, one quarter, one success metric agreed in advance. Small enough that the cost sits inside your boss’s own discretion, and bounded enough that a no in month three is a data point instead of an embarrassment.

Anchor it against the cost of doing nothing. A pilot that automates one manual reporting process might cost $15,000 to $25,000 all in. The status quo is not free: that same process running one more year at half of a loaded full-time salary is $40,000 to $60,000, spent invisibly. Put both numbers in the same sentence and the pilot stops looking like an expense.

What you are not asking for yet: a company-wide AI strategy, a platform decision, or headcount. Those come after the pilot gives you real numbers to argue with.