Experts Advise How to Get Better Results from AI by Improving User Prompts

Many users express disappointment with artificial intelligence tools because they treat AI like a search engine, inputting short keyword phrases instead of detailed prompts. Unlike search engines that retrieve existing links, AI language models generate new responses based on the context and information provided. The more detailed and specific the input, the more accurate and useful the output will be.

Noam Cohen, founder of CritiqeIL, a consultancy specializing in AI implementation in Israeli organizations, outlines five practical steps to improve AI interactions. First, provide context instead of just keywords, such as specifying the audience, tone, product, and goal when requesting content ideas. Second, assign a role to the AI, like "act as a marketing consultant," to guide tone and perspective. Third, include examples of desired style rather than describing it verbally. Fourth, specify the format of the response, such as a list or table, to receive ready-to-use outputs. Finally, treat the first AI response as a starting point and continue refining it by asking for edits or additions.

Cohen emphasizes that the key difference between users who are disappointed and those who benefit from AI is not the tool itself but how they communicate with it. He encourages managers to train their teams to engage with AI conversationally rather than as a search engine. Patience and iterative dialogue with AI can significantly improve results without needing more expensive tools. Cohen suggests users try expanding a short three-word query into a full paragraph to see the difference in AI responses.

This guidance aims to help users unlock the full potential of AI by shifting their approach from brief keyword searches to detailed, interactive conversations with the technology.