Every other AI product on the market is trained to keep talking. If it does not know, it guesses. If it has nothing to draw on, it invents. The product is judged on confidence and fluency, not accuracy.

Ours is built the opposite way. When the library does not cover your question, the chat says so out loud. That refusal is not a bug. It is the most important feature we have.

What a refusal looks like

You will see something like this:

"Our library does not have enough on this to give you a useful answer. For a question like this, you would want to talk to a CPA who works with small operators in your state."

It is short, it is honest, and it points you somewhere better. It does not invent a tax code. It does not paraphrase a Reddit thread. It does not make up a statistic.

Why we built it this way

Owner operators do not need another tool that sounds smart. They need a tool they can trust with a real decision. The cost of one hallucinated overtime rule, one made up warranty obligation, or one fabricated tax deadline is bigger than the cost of being told no a few times.

How to read the refusal as a signal

A refusal tells you three things at once.

  1. The question is outside what we can answer well today.
  2. The answer probably belongs with a licensed professional or a peer with direct experience.
  3. The question is now on our roadmap. We log every refusal so we know what to expand next.

How to get a better answer when you hit a refusal

  • Tighten the question. "Pricing strategy" is broad. "Pricing a roof repair for a 1,900 square foot ranch in central Texas" is specific.
  • Give us the context. Your trade, your market, your team size, and the version of the problem on your desk.
  • If we still cannot answer, the refusal will name where to go next.

The broader point

"I do not know" is what an honest peer says when you ask them something past their experience. We built the chat to act like that peer. Less reach, more trust. That trade off is the whole product.

For the longer take on why we drew the wall where we did, see why we built a closed wall instead of using the open web.