AI: Ask for Options, Not Answers

Posted on July 28, 2026

One of the simplest shifts you can make to improve results when using AI is to avoid asking the AI tool for a single answer. Ask for four or five scenarios instead. Do that, and you’ve got raw material to combine with the particulars of your situation that you already know. This approach helps improve the deliverables you produce for your projects. Here are other reasons why asking for options is so effective:

  • A single answer strips out the context only you have. When facing a client who’s asking a challenging question, I’ll ask the AI for four or five scenarios describing how others have handled similar situations. This directs the AI to supply patterns. I can then assess suitability and enhance my deliverables accordingly.
  • Forcing multiple options makes the AI do more analytical work. Asking for options rather than a single output pushes the tool to consider more angles rather than converging on the first plausible-sounding response. You end up with richer material to work from.
  • Asking for options can launch a much richer “discussion” with the AI. A university course I designed requires students to use an AI in an assignment. I’ve asked them to show their original prompt, the first response, three follow-up prompts that improve it, and what they’d still refine afterward, whether or not they use AI to do so. The responses I have seen that start with options rather than a single answer result in much richer discussions and better deliverables.
  • Choosing between options is the human part of the job. Being handed five ways to phrase something is only useful if you can tell which one fits this person, this moment, this relationship. That judgment call is exactly what the AI can’t produce. (At least not yet, without a lot of impractical training of the AI.)
  • This mindset also protects you against getting duller. A colleague and I had a conversation in which we were both concerned that AI is making us worse writers. We concluded that querying for multiple sources of raw material to shape, rather than asking for a fully written product, keeps us doing the thinking instead of outsourcing it.