Getting useful results from your first AI chats

A useful first exercise is a task you already understand. You’ll have a better sense of what a good answer looks like and where the AI misses something. Try planning a community event, outlining a presentation, or improving a piece of writing.

Choose something small

Suppose you’re helping a Frederick community group plan an introductory technology session. You need an agenda, not a complete event-management system. Start with invented details or information that is already public.

Write down what you want to have at the end: a 45-minute agenda for beginners, with one exercise and time for questions. That gives you something specific to evaluate.

Give the chat enough context

Explain the task, who it is for, the constraints, and the format you want. Microsoft’s prompting guidance similarly emphasizes goals, context, expectations, and source information.

This example provides boundaries without prescribing every detail. You can adapt it to a team meeting, a lesson, or another planning task.

Use the next message to improve it

Read the first response as a draft. If the exercise assumes everyone has a laptop, tell the chat that people will be sharing one screen. If the agenda is too technical, identify the section that needs simplifying.

Try a precise follow-up: “Make the exercise work with one shared screen, and show the revised timings.” Keep useful parts and ask for changes to the rest.

Check what matters

For this agenda, add up the minutes, check that the activity fits the audience, and verify any factual claims against reliable sources. If the chat supplies links, open them and see whether they support the answer. Microsoft also recommends checking AI output against your source material in its output review guidance.

Save the final prompt, your corrections, and the result. That small record gives you a starting point for the next task—and helps you see where your own judgment made the difference.