Improve Your AI Results by Writing Better Prompts

Questions

  • How do I write better prompts for AI tools?
  • What is prompt engineering and why does it matter?
  • How can I get more useful answers from ChatGPT, Copilot, or LibreChat?
  • Why is the AI giving me vague or unhelpful responses?
  • What tips or examples help improve AI-generated results?
  • How do I ask AI tools for step-by-step explanations?
  • Where can I learn more about writing effective AI prompts?

Environment

This article is for all Bowdoin College students, faculty, and staff who use AI-powered tools — including LibreChat, Microsoft Copilot, Google Gemini, or Bowdoin Bot. The techniques apply to any conversational AI system. No special tools or software are required beyond access to an AI tool.

Tip: Bowdoin Bot, the chat assistant available on every Knowledge Base page, is a generative AI bot. The prompt techniques in this article will improve your results with Bowdoin Bot as well.

Resolution

What is prompt engineering?

Prompt engineering is the practice of crafting clear, specific instructions — called prompts — for AI systems. A well-written prompt helps the AI understand exactly what you need and produce a more accurate, useful response. You do not need technical skills to benefit from it.

Techniques for better prompts

Be specific about what you need

Vague prompts produce vague answers. Include the context, the format you want, and any constraints that matter.

  • Less effective: "Tell me about climate change."
  • More effective: "Summarize the three main causes of climate change in plain language, suitable for a first-year college student."

Specify the format or output type

Tell the AI how you want the response structured — a bullet list, a table, a paragraph, numbered steps, a formal tone, or a casual explanation.

  • "Summarize this article in three short bullet points."
  • "Write this as a formal email to a supervisor."
  • "Give me a step-by-step guide with numbered steps."

Ask for follow-up questions

Many AI models can ask clarifying questions before answering. If you are not sure how to specify your request, try asking the AI to ask you for more details first.

  • "Before answering, ask me any follow-up questions you need to give me the best response."

Break down complex requests

For multi-part questions or large tasks, split the work into smaller steps rather than asking for everything at once.

  • "First, list the main causes of climate change. Then, in a follow-up, explain one of them in detail."

Provide context or examples

If you want the AI to match a style or build on existing work, paste in a sample or describe the context directly.

  • "Here is the first paragraph I wrote. Write the next two paragraphs in the same tone."
  • "I am writing for a non-technical audience. Avoid jargon."

Refine and iterate

If the first response is not what you needed, do not start over — ask the AI to adjust. You can ask it to be shorter, more formal, more specific, or to focus on a different angle.

  • "That was too long. Rewrite it in two sentences."
  • "Focus more on the financial impact and less on the historical background."

Additional Help

If you need further assistance, you have several options:

  • Bowdoin Bot: Chat with Bowdoin Bot directly from any KB page for instant answers.
  • Phone: Call the Bowdoin College Service Desk at (207) 725-3030.
  • In person: Visit the Tech Hub in Smith Union during business hours.
  • Submit a ticket: Request assistance through the Service Catalog.

Additional Resources

Print Article

Related Articles (4)

Instructions for accessing LibreChat, Bowdoin's web-based AI chat tool that provides access to multiple AI models including OpenAI GPT and Anthropic Claude. Covers how to sign in, choose a model, and use the tool safely. Applies to all Bowdoin students, faculty, and staff.
An overview of the three main AI tools available to Bowdoin students, faculty, and staff — Microsoft Copilot, Google Gemini, and LibreChat — including how to sign in to each and basic guidelines for responsible use.
Instructions for installing LM Studio on a Bowdoin-issued or personal Mac and using it to download and run AI language models locally, without sending data to external servers. Covers installation via Jamf Self Service, downloading a model, starting a chat, and keeping LM Studio up to date.
Instructions for using Bowdoin Bot, the AI-powered chat assistant available on every Knowledge Base page. Covers how to open the chat, ask questions effectively, and what to do if the bot cannot find an answer.