What is the difference between Agenframe, a regular prompt, and Custom Agents?

What is the difference between Agenframe, a regular prompt, and Custom Agents?

AgenFrame™
Regular Prompt (non-agentic)
Custom Agent

Agenframe is used for more complex queries involving multiple tasks. It produces structured output responses with bullets and other lists.  It provides visual audit trails, explanation of how agents compile results and can produce in-depth research results ("InsightApex"). AgenFrame is not custom and produces a self-organizing team of dynamic agents that execute a user's desired goals.  

Regular is used for general prompts, or when a quick answer is needed. Regular prompts can execute tasks and multiple steps but may not produce consistent results as data sets change. They use memory within a conversation for follow-on queries when a user states in the "previous response" or mentions previous. 

Custom agent response involves using an agent which has been custom built for the user to best response to a specific type of question. Custom agents have pre-defined steps and also have access to custom tools which have been built specifically for the user.

Use cases for specific prompts:

AgenFrame: AgenFrame prompts are suitable for questions that cannot be answered by a single retrieval or simple synthesis. This includes queries requiring reasoning across different pieces of information, combining data points, following a logical chain of thought, or generating a response that builds upon intermediate findings. The dynamic query planning makes it flexible for handling a range of complex questions, without needing to spend time to create a custom agent for a specific task.

Regular Prompt:  Regular prompts are best suited for simple, factual questions where the answer exists directly within the ingested documents and can be extracted or synthesized from retrieved snippets without complex reasoning or external tool interactions. For example, asking for a definition, a specific piece of information contained within a document, or a summary of a topic covered in the data

Custom Agent: Users create Custom Agents to solve specific problems, generate specific items a client needs, or handle complex tasks. This means if a standard query or even a multi-step AgenFrame™ isn't producing the exact desired output or performing the required action, a Custom Agent can be built to ensure that specific outcome. Examples include generating complex reports, answering questions with a specified format, or comparing documents

    • Related Articles

    • How can I rerun a prompt in Odyssey?

      Standard Prompts: If you're looking for a more comprehensive response from a standard prompt (without AgenFrame), simply click the 'Rerun' button at the end of the response. You’ll get a new result in just a few seconds. AgenFrame and Custom Agent ...

    • What are the key words to trigger an accurate response to my previous prompt result?

      Odyssey AI allows for up to 3 additional follow-on questions to your original prompt query. This ensures that follow-on responses are contextually relevant, accurate and the source citations (SourceSnippets) stay correct. Use any of the three phrases ...

    • How to save prompts to the prompt library?

      At the end of every prompt result, use the bookmark to save your prompt. Your prompt results is private and only available in your workspace.

    • Release Notes April 2025

      Odyssey AI UI Enhancements Significant improvements have been made to Odyssey AI’s user interface to enhance your experience. Response Sharing: You can now share Odyssey AI responses with any team member in just a few clicks. Note: Sharing is only ...

    • AgenFrame™ & Agentic XAI™

      <br>