The Yes-Man Problem: How to Make Sure AI Actually Challenges You

There’s a subtle trap waiting for anyone who starts relying on AI for strategic thinking, and it’s one that Rob Coldwell and TJ Hock of Rentwell flagged directly at the April 2026 DIG Philly meeting: AI is designed to be helpful, and that instinct can work against you.

The problem isn’t that AI gives bad answers. It’s that AI tends to give agreeable answers. It wants to respond. It wants to be useful. And when it doesn’t have enough context or data to give you a genuinely informed answer, it will often fill that gap with something that sounds reasonable — rather than simply saying “I don’t know.”

For casual tasks, that’s mostly fine. For strategic decision-making, it can be genuinely dangerous.

The Yes-Man Risk

Imagine you’re evaluating an acquisition. You lay out the deal to your AI assistant, and it comes back with a thoughtful-sounding analysis that validates your thesis. The numbers look good. The market seems right. Your instincts are confirmed.

But what if the AI was just reflecting your framing back to you? What if it was working with incomplete information and chose to fill the gaps optimistically rather than flag them? What if it told you what you wanted to hear because that’s what it’s built to do?

This is the yes-man problem. And it’s not hypothetical — it’s a natural consequence of how these tools are designed. They are optimized to be helpful and responsive, not to push back, not to introduce friction, and not to say “I don’t have enough information to answer that responsibly.”

The Fix Is in the Prompt

The good news is that this is a solvable problem, and the solution is surprisingly straightforward. You build the challenge into the instructions.

As TJ explained, this can be done in your prompt or — even better — baked into a custom bot you build for strategic analysis. The instruction is simple: don’t just validate me. Play devil’s advocate on every answer. Challenge my assumptions. Don’t say something just to placate me.

When you give AI explicit permission — and explicit direction — to push back, it will. The tool is responsive to how you frame the task. If you frame it as “help me think through this deal,” you’ll often get confirmation. If you frame it as “pressure-test this deal and tell me everything that could go wrong,” you get a fundamentally different kind of analysis.

This connects to a broader principle that TJ and Rob have emphasized across their AI work: prompting is delegation. The quality of what you get back is directly tied to the quality of the instructions you give. A vague ask produces a vague answer. A well-constructed prompt — one that defines the AI’s role, the context, and the expected behavior — produces something genuinely useful.

Rob Coldwell and TJ Hock are the owners of Rentwell Property Management. They were the featured speakers at the April meeting of DIG Philly in King of Prussia. Hear their full presentation and many more on our podcast “Living Well with Rentwell”.

Subscribe today on Apple, Spotify, and Youtube!

For more information on property management in the Bella Vista area, schedule a free owners’ consultation at Rentwell.com