Every client we work with has the same internal clash, and we heard each side of the argument on two separate calls this week:

“I’ve got executives that have their head in the sand, saying 'nah, we're a widget company, we don't need AI, this is bullshit, all it does is hallucinate.’”

- CEO, PE-Backed Widget Company

“A lot of us are dinosaurs who remember doing things the way we did, and we're excited to figure out how to shift gears. We don't know where to start but know we need to.”

- President, PE-Backed Automotive Services Company

The AI enthusiasts push three-steps ahead and still think the business is moving glacially slow. The AI skeptics protect the existing best practices from devolving. Neither side is right or wrong, but both come with risks that must be recognized:

1. Enthusiasts: the AI power users move fast, test ideas, and find AI use cases across the business. They risk burning time and budget on projects that look impressive but change nothing. They'll lose patience and motivation when asked to slow down and prove it.

2. Skeptics: the cautious AI users or abstainers shield the company from expensive nonsense and won’t budge until it’s worth moving for. They risk inhibiting progress.

Each side depends on the other to regulate their own risks. When the skeptics dominate, no changes get made. When enthusiasts prevail, rapid “progress” goes unchecked. We’ve led some feisty client calls, but, much to the AI enthusiasts’s disappointment, we don’t take sides.

Instead of right and wrong, we’ve found that it’s better to facilitate a healthy conversation between both sides.

We met a CEO who described his Head of Sales as their “prized AI power user.” It turned out the guy was building a new agent every month that made the prior one redundant. Quantity sacrificed quality, and they would have benefited from some skeptics in the room.

The teams with the best tension, seem to steer each group to the right questions about AI. 

We met with a technology company whose VP of Engineering was concerned they were using too many tokens. A three thousand dollars Claude bill felt egregious until his boss, the CTO, asked what work got completed. It turned out the VP had developed a year's worth of output. It became clear to the full team, including the skeptical VP, that he should push the AI gas pedal down.

The skeptic was right to be worried about costs. However, he needed to know what the cost resulted in before he could trust it.

So here’s the game plan: 

• List out the AI projects in your business from the past year and designate each one enthusiast-led or skeptic-blocked. Enthusiast projects need one number proving they work, or they get paused. Skeptic projects need one number showing what waiting is costing, or they go back in play.

Keep your skeptics doubting and your enthusiasts buzzing. But you need each of them answering the questions that disarm the risks that they bring to the table.