Skip to main content

​
What Is a Prompt?

A prompt is the instruction, question, or text you provide to an AI model to get a response. It’s how you communicate with a Large Language Model (LLM).
The quality of your prompt directly impacts the quality of the answer you receive.
Think of it this way:
  • Poor Prompt: “Tell me about marketing.”
  • Good Prompt: “Summarize the three most important marketing strategies for a B2B software company in a bulleted list.”

​
How Prompts and LLMs Work Together

LLMs use your prompt as a starting point and guide. Based on the instructions you provide, the model predicts the most logical and relevant sequence of words to form an answer. A clear and detailed prompt helps the LLM understand the precise context and your expectations, leading to a much more accurate and useful output. Read our guide about LLMs to find out more.

​
Find Out How to Use Genow and Prompt Optimally

Different features within Genow are designed for specific tasks. Understanding how to interact with each one will help you get the best results. Comparison between the Use Case Hub and the Chat function and how this affects prompting.
CharacteristicUse Case HubChat
Intended UseExtracting precise answers from company internal knowledge bases and executing domain-specific workflows.General-purpose assistance for writing, reasoning, and creative brainstorming.
BenefitsTime-saving, highly accurate knowledge retrieval across complex and large amounts of internal data.Versatile support in daily work to save time, reason and explore new ideas.
Data ResourcesCurated, pre-configured corporate datasets and managed web search. Hallucination of models is prevented.Pre-trained model knowledge and user-provided manual attachments. Hallucinations may occur frequently.
CustomizabilityGreat customization within use cases through data integration, web search, contact persons, links, etc.Limited customization only through information provided in the prompt.
TransparencyDirect citations and links to source documents and outline of the reasoning process.Responses are generated from internal knowledge without source citations.
Prompting StyleFor a new question open a new thread. Separation between retrieval and task: 1) Define a precise question stating what is sought-after and the context. The whole prompt is used for semantic search. 2) Based on the retrieval results, define a clear task describing the expected outcome.Interactive, creative and versatile prompts that are specific, provide context, give examples and define the output.
The Use Case Hub is optimized for finding answers and information within your company’s dedicated knowledge base in a use case. Generally you can follow a guideline to use the Use Case Hub optimally:
1

Internal Document Search

Select a use case and perform a detailed search based on the documents and knowledge assigned to the use case.
2

Perform Tasks based on the answers

Use the agents and actions feature to perfom task (“write an e-mail draft”, “Convert the results into the following format: …”)

Learn more about the user interface and usage of the Use Case Hub.
Best Practices:
  • Be Specific: Ask detailed questions about the knowledge stored in the use case.
  • Define The Output: Tell the model how you want the answer formatted. Do you need a summary, a bulleted list, or a table?
  • Provide Context: Briefly mention the context of your question to get a more relevant answer.
  • Start A New Thread For A New Topic: Always start a new conversation for a new question. Using an existing thread can confuse the AI with a previous conversation’s context, which may negatively affect your results.