How is AI like a pilot lost above the clouds near Redmond, Washington?
A technically correct answer can still be worthless.
When companies pay for the tokens that produce it—and the employee time spent trying again—that’s an ROI problem.
Poor prompts contribute to that waste. But why should users bear all the responsibility for getting the question right?
That’s why I built Collaborator™.
There’s an old joke about a helicopter pilot who gets lost near Redmond, Washington, after his navigation and communications equipment fail.
Through a break in the clouds, he spots people in a tall, unmarked office building. He flies close and holds up a sign:
WHERE AM I?
The people inside confer, then hold up their answer:
YOU ARE IN A HELICOPTER!
The pilot immediately turns toward an airport and lands safely.
“How did that answer tell you where we were?” his passenger asks.
“It was technically correct—but completely useless. So I knew it had to come from Microsoft.”
Sometimes, using AI feels like that.
We give an LLM an instruction. It delivers a polished, technically correct answer—without first establishing what we’re actually trying to accomplish.
Proficient AI users compensate. They provide context, challenge assumptions, recognize misunderstandings and keep steering toward a useful result.
But most people aren’t proficient AI users. Nor should they need to be.
The user shouldn’t need to know how to prompt the AI.
The AI should know how to prompt the user.
That’s the premise behind +Collaborate™.
I built a set of additional system instructions designed to help AI discover the user’s desired outcome and gather the information it needs to help achieve it.
When a request is incomplete, +Collaborate™ can ask purposeful questions, surface assumptions and draw on the user’s own knowledge, judgment and experience.
I’ve built an interactive demo so you can compare the same model with and without Collaborate™. Choose a suggested prompt or bring one of your own.