0 min read

Interview with AI

Prompt Engineering

Context Engineering

Most AI workflows start the same way. You stare at a blank prompt box, try to describe everything you need in one shot, hit submit, and spend the next hour correcting assumptions you never asked the tool to make. The coffee gets cold. The revision count does not.

There's a different way to start. Describe the rough idea and tell the AI to ask you every question it needs before doing anything. Don't guess. Don't assume. Just ask.

The dynamic shifts immediately. Instead of trying to front-load every constraint and requirement into a single prompt, you're being interviewed by something that won't proceed until it actually understands the problem. The output that comes after is almost unrecognizable compared to what a cold prompt produces.

The reason many people skip this: it takes longer. There are more steps, more back and forth, more thinking required upfront. That friction is exactly the point. The assumptions that don't get surfaced at the start don't disappear - they show up later as wrong outputs, wasted iterations, and prompts that start with "no, that's not what I meant."

Good design work has never had a version that skips the thinking, and AI didn't create one.

Alex Dihel | Product & Marketing Design Leader | Design Operations   www.alexdihel.com © | Privacy

Alex Dihel | Product & Marketing Design Leader | Design Operations   www.alexdihel.com © | Privacy

Alex Dihel | Product & Marketing Design Leader | Design Operations   www.alexdihel.com © | Privacy