Understand AI reasoning¶
Alan AI offers visual reasoning graphs — flowchart-like diagrams generated for every response the AI agent produces. Reasoning graphs offer a detailed view of the entire AI reasoning process and make it easier to understand how the AI arrives at its conclusions. With reasoning graphs, you gain insight into the AI decision-making, better understand how the AI agent operates and refine the AI agent behavior.
Use case¶
For an AI agent developer, it is essential to understand the AI reasoning process. This includes:
The data sources the AI agent uses to answer user queries
How the AI agent decides which actions to take based on input data and dialog context
The code it generates to retrieve the required data
How the AI agent constructs its responses
Prerequisites¶
To successfully follow this exercise, make sure the following prerequisites are met:
You have signed up for Alan AI Studio and created a project for the AI agent. For details, see Sign up for Alan AI Studio.
You have set up static and dynamic corpuses for the AI agent. For details, see Create a static corpus and Create a dynamic corpus.
Review the reasoning graphs¶
To understand the AI reasoning process:
In the Debugging Chat, ask several questions to query static and dynamic corpuses, for example:
What are capacity levels for local SSD?Show available buckets
In the lower part of Alan AI Studio, open the Alan AI Studio logs and click the Chart icon next to the AI agent response.
Review the reasoning graph. As part of this process, you can:
See the general flow of the AI reasoning process.
Drill down into a specific reasoning step. To do this, click the necessary block in the graph.
Understand what data the generated code returns: to do this, in the Generated code block, click the Run button in the top right corner of the window.
Review the output format: to do this, in the Output block, switch between data display modes using buttons in the top right corner of the window.
