Agent Approaches

We always recommend that an Agent has a specific purpose and conducts a specific Task over specific Knowledge. This is to ensure the output quality from your Agent is as high as possible - conflating different purposes and tasks within a single Agent increases the risk of the Agent "getting it wrong".

Here is our Standard Agent:

Here are some Agent Approaches for inspiration:

Approach	Description
Categorisation	Classifies retrieved data into relevant groups or labels, helping users make sense of complex information.
Structuring	Organizes retrieved information into structured formats like tables or templates, making it easier to act on.
Refinement	Refines retrieved information by filtering, rephrasing, or reorganizing it to make it clearer and more actionable.
Question-Answering	Uses retrieved information to answer user queries with precise, evidence-based responses.
Knowledge Validation	Cross-references retrieved information with multiple sources to confirm accuracy, reliability, and relevance.
Comparison	Uses retrieved information to create comparisons between scenarios, options, or strategies for decision-making.
Conversational Guidance	Engages the user in dialogue to clarify their intent, refine their requests, or ensure the input is well-structured before proceeding.
Input Validation	Checks the quality, format, and relevance of user input and suggests corrections or improvements to optimize the outcome.