Getting Better Results from Generative AI: 9 Practical Tips

Let’s start with a basic truth: most people don’t understand how AI works, and many distrust the answers generative IAs give. Ironically, those who do use generative AI often fail to verify the results. Some are intimidated by it, while others misuse it in ways that increase the risk of getting misleading or wrong results.

Based on my own findings—plus insights from Harvard University IT and common best practices shared across reputable sources — I’ve put together nine of my favorite tips for getting better results from generative AI.

Five Ways to Be Specific

Generative AI draws from the entire known universe of content, so you must narrow it down. The more specific and structured your prompt, the better your results.

1. Provide relevant context and background

   Don’t assume the AI knows what you're talking about. Briefly describe how you arrived at the question in mind and what you're trying to achieve. Frame the issue clearly.

   
   

2. Break things down

   Complex queries require exploration and discovery and can't be divined from a single shining question. Break down your query into it's logical parts. For example, if you're researching a car to buy, ask about safety, fuel economy, and reliability as separate questions — then refine.

   
   

   Second, avoid feeding the AI all the research steps at once because it will attempt to give you all the answers at once without building up to the best result. Feed it one step at a time instead.

   
   

3. Clearly state your desired output

   It's important to tell the AI the form of answer you will be looking for. If you want cowboy poetry, say so. If you want a single best recommendation, or if you want a list of possible arguments to consider, make it clear.

   
   

4. Be extra specific to avoid confusion

   Review your query again for any possible ambiguities. If there's more than one possible way to take a phrase, the AI will pick one possibility out of a all the rest, and chances are it won't be the one you wanted.

   
   

5. Define scope to set boundaries

   Try to define the boundaries of what you want — how far, how much, how deep. Most especially, tell the AI what you don't want, so that it can avoid unnecessary tangeants.

Remember: It's just a machine.

The more you use AI, the better you can understand that the large language model (LLM) it uses to communicate is not at all as human as it sounds and needs to be treated as simply a logical device, incapable of real independent thought.

6. Be aware of the limitations of the AI

   AI only simulates human responses, and it doesn't think. It has no ethics, no values, no real emotion or emotional intelligence, so it can't 'read between the lines' or understand social implications. It simulates creativity, but it's not really creative. It mimics problem solving based on patterns in data. And while it’s usually polite, be mindful that it can produce answers with unintended tone or bias.

   
   

7. Structure your prompts for clarity

   It helps both you and the AI to structure your request using bulleted lists, templates, or formatting. Your AI can suggest a prompt template. A well-structured prompt often produces a well-structured response.

   
   

8. Verify your results

   I found it interesting that none of the AIs I asked suggested this item (ahem). In the spirit of "trust, but verify" it's always good practice to carefully review the AI's response. Ask it for links to prove it's results, then follow the links. I am often surprised at the assumptions AIs will make when generating a response.

   
   

9. Iterate: make corrections and give feedback

   Lastly, getting good content from an AI usually takes multiple rounds. Don't give up on the first try and don't take anything at face value. Get your your (virtual) red pencil and make all the corrections you can find. Then review the above eight steps and try again. After each try, you should feel the AI getting closer to providing the answers you need.

   
   

Final Thought:

The key to using AI well isn’t technical prowess — it’s intentionality. The more thought you put into the prompt, the more useful the result. It's a powerful tool, but ironically it can't think for itself. When it keeps changing your drafts, consider the changes, but you don't have to accept them. You are the boss!