If artificial intelligence is part of your product—and judging by the startups I’ve seen recently, it probably is—you have a lurking messaging challenge.

Even without AI you may have an analogous issue if a successful sale requires the buy-in of multiple stakeholders.

I first encountered it when Michael Murphy from Prelude hired me a few months ago to solve a frustrating problem: He would speak with a potential client CEO about Prelude’s AI-driven system that optimizes and automates inventory management for cannabis dispensaries. “Take the guesswork out of purchasing,” is how Michael summarized his pitch in calls and on his website. “Do in minutes what previously took weeks.”

The CEO would love it.

“Then the CEO would talk to his staff,” Michael told me. “And the sales process would grind to a halt.”

Interesting mystery. I spoke to dispensary workers and industry veterans. It became clear—obvious, even, once I saw it. While there was nothing factually incorrect with how Michael framed his value proposition, the actual buyers at the dispensaries had two clear reasons to hate this product pitch: first, it offended them by implying their current approach was “guesswork.” Second, to them, ‘efficiency’ sounded like another word for ‘unemployed.’

I’ve since seen this same pattern at other AI startups, including, just in the past few weeks, one for fashion designers (met via 500), manufacturing inventory management (also 500), and another for restaurant owners (met via PAX).

“Intelligent automation” is the essence of the AI value proposition. But be careful: What sounds great to the CEO and CFO might land completely different with affected managers or departments. No one ever supported a product suggesting they are ‘dumb and dispensable.’

Instead, you need a pitch that addresses the hopes and fears of all stakeholders.

How do you do that?

Painting that picture to the CEO is easy: more success, more profit.

To those on the receiving end of your “intelligent automation,” it’s more subtle.

First, don’t disparage the way work is currently done. Claims to eliminate “guesswork,” “chaos,” or “inefficiency,” as accurate as they may be, sound like accusations of incompetence. Phrases like “discover patterns hidden in your data,” accomplishes much the same without the accusatory tone.

Second, highlight that AI automates tasks, not jobs. Most jobs are a bundle of tasks. Make it clear that the AI handles the hard, boring, complicated parts of a larger job, making time to focus on the more fun, fulfilling, and ultimately higher value-add tasks of that job. Saving time is good, in this framing, because it gives you ‘freedom to focus on the good stuff.’

Imagine AI as an espresso machine, and you’re pitching it to a barista currently making each order from scratch. Automation sounds like unemployment. But imagine a coffee shop that already has espresso machines. They still have baristas, and they’re still busy. But instead of focusing on the mechanics of production, they’re interacting with customers, making fancy foam art, and selling more pastries. They’re doing more fun, higher value work.

Don’t pitch what goes away. Highlight what gets enabled.

I rewrote the Prelude site. The CEO gets “fresh product, happier customers, healthier bottom line.” The buyer hears they can “unload the boring tasks that clog your day so you can focus on the job you were hired for.”

Michael has found much better resonance with the new messaging. It’s not a silver bullet, he says, but it eliminated a key sales roadblock.

Are you applying human intelligence to your artificial intelligence pitch?