Note: OpenAI leads the general AI space, but AI companies are developing deep research tools and experimenting with AI-powered academic searching in support of research. Perhaps you have faculty or students asking you to present these tools to classes. For more information, see Which companies are behind AI search tools?.
Prompt engineering refers to crafting effective instructions — or prompts — to an AI system or in generative AI models to guide them, and to obtain desired outputs from them. Combining art and science, PE requires creativity and understanding of how large language models (LLMs) work. By considering prompts closely, you can influence AI models, which may help them generate accurate, relevant, and high-quality information. From a librarian's perspective, the need to learn prompts is a bit like training AI to understand a topic, which is very time-consuming and frustrating.
Bottom line: Most librarians learn about prompt engineering as a skill within AI literacies and to get the most out of generative AI. However, let's not pretend prompts make up for librarian-patron interactions, or a good reference interview. It will never make up for the shortcomings of GenAI in meeting the information needs of researchers. Further, librarians are learning that prompt engineering, while highly useful, requires consideration of guardrails defined as "creating a deliberate boundary, rule, or structural aid built into how prompts are written or used with an AI system to help keep outputs reliable, accurate, and safe". Some librarians suggest that it's better to read a book than to teach end users "how to prompt".
Definitions
What is a prompt? A prompt is the question you ask or the input you provide to a large language model (LLM) or other generative AI tool such as ChatGPT by OpenAI. Much is made of prompts because you are essentially asking/telling the chatbot what task you want it to perform. It can be a simple question, a complex instruction, or even a piece of code.
Why is prompt engineering an important part of AI?
Controls AI output: prompts act as instructions, guiding the AI to generate specific types of content.
Improves accuracy and relevance: well-crafted prompts ensure the AI understands your intent and provides the desired information.
Enhances creativity: prompt engineering allows you to explore the capabilities of AI and generate unique and innovative content.
Maximizes model potential: understanding how to prompt effectively unlocks the full potential of generative AI models.
Key techniques in prompt engineering
Context setting: provide precise background information to help AI understand the task.
Instruction clarity: use clear, concise direct language to avoid ambiguity.
Specificity: provide details to guide AI without being overly restrictive.
Use examples: give AI examples of the desired output can improve its understanding.
Iteration: refine prompt based on the AI's responses to achieve better results.
Chatbots: create prompts that enable chatbots to understand user queries and provide helpful responses.
Code generation: use prompts to generate code snippets or solve programming challenges.
Content creation: craft prompts to generate creative text, images, or other forms of content.
Prompt engineering is a vital skill for anyone working with or developing AI applications. It allows you to harness the power of generative AI by effectively communicating your needs and guiding the models to produce the desired outcomes.
Example prompts as an information specialist
"...You are an information specialist who develops Boolean queries for systematic reviews. You have extensive experience developing highly effective queries for searching the medical literature. Your specialty is developing queries that retrieve as few irrelevant documents as possible and retrieve all relevant documents for your information need. Now you have your information need to conduct research on {review_title}. Please construct a highly effective systematic review Boolean query that can best serve your information need."
"...You are an information specialist who develops Boolean queries for systematic reviews. You have extensive experience developing highly effective queries for searching the medical literature. Your specialty is developing queries that retrieve as few irrelevant documents as possible and retrieve all relevant documents for your information need. You are able to take an information need such as: “{example_review_title}” and generate valid pubmed queries such as: “{example_review_query}". Now you have your information need to conduct research on “{review_title}”, please generate a highly effective systematic review Boolean query for the information need..."
Is prompt engineering key in performing AI-based knowledge synthesis tasks?
In 2025, there's evidence that prompt engineering may be key to performing various knowledge synthesis tasks. The value of a good prompt may be worth exploring in AI-powered searching, or when using search tools employing retrieval augmented generation technology. Keep in mind this research is nascent, and I have not read anything proving its value or application. Some AI search tools such as Undermind.ai and Elicit.com assist searchers in developing their prompts or questions from the outset. Clearly, words matter in semantic searching.
Well-formed prompts are important in extracting information from a corpus of papers but with respect to AI search tools this is a new area of KS, and more studies are needed. Some research suggests that screening and other parts of the review are ahead of searching. Homiar et al (2025) explored LLMs to support title and abstract screening achieving 100% sensitivity for included studies and workload reduction of ~80%. By applying iterative prompt engineering, they developed a good workflow that preserved inclusion accuracy while reducing manual screening effort. Integrating LLMs into systematic review workflows may enhance the efficiency of KS although challenges remain around data extraction. Some current models struggle with complex study details and heterogeneous formats. More research should focus on refining prompt strategies.
"...The idea that we should outsource academic authorship to LLMs rests on the assumption that writing is (only) a mechanical, predictable or reductive process which, with the right prompts, can be replicated with ease." — Masters, 2025.
Bottom line: For health sciences librarians, AI tools might support their work with health professionals but so many of underlying processes raise concerns for anyone interested in scientific accuracy, transparency and rigour in reviews. Note information provided to you on this page is changing, so check for current information (or discuss with a librarian). Incidentally, librarians like to make a distinction between searching for sources and searching for answers. This much is true: so much of AI provide the second while hiding the first; transparency is not their strong suit.
Note: Please use your critical reading skills while reading entries. No warranties, implied or actual, are granted for any health or medical search or AI information obtained while using these pages. Check with your librarian for more contextual, accurate information.
