Advanced Settings

Advanced

Response settings

Show Tech details

Toggle to show or hide special block with the chain of thought. This allows you to understand the thinking behind the AI Analyst’s answer, which may be useful for reasoning and testing purposes.

Show Assets

Toggle to show or hide special block with the list of assets (datasets, tables, charts) used to create the response.

Always show table as output

This setting controls whether tables are always included in the AI Analyst's responses.

AI Analyst behavior

Show Consistency Check warnings

AI Analyst will warn you in case prompt was identified as not consistent with the available datasets and context.

Show Question Confirmations

This feature enables the AI Analyst to confirm its understanding of the user's prompt by rephrasing the question and seeking user approval before proceeding with the response.

Use previous responses as a context

When enabled, the AI can use its previous responses as part of the conversation context, but only if pipelines are executed.

Performance

Preferred LLM

Select between two LLM models for the AI analyst, preferring either response speed or response accuracy.

Context Memory Span

Choose how long the AI Analyst should remember interactions. Short retains context for the current session, Medium for a few sessions, Long for extended periods, and Unlimited indefinitely.

Privacy

Send Categories to OpenAI

Toggle to enable or disable the sending of categorized data to OpenAI. When enabled, the AI will send predefined categories along with your queries to OpenAI, allowing for more context-aware and relevant responses.

Generate Summary Using OpenAI

Toggle to enable or disable the generation of summaries using OpenAI. When enabled, the AI will use OpenAI's capabilities to generate concise and coherent summaries of the provided data or responses.

Use numbers in summaries generated by OpenAI

This setting controls whether the AI Analyst includes actual data in the text part of the response.

When enabled, numerical data will be incorporated into the summaries generated by OpenAI. Note that all data is securely sent to OpenAI in an obfuscated manner to ensure privacy and security.

Testing and Monitoring

Langsmith Project Id

This ID helps in organizing and managing your AI Analyst's logs and evaluations within the correct project. You can find this ID in your Langsmith dashboard under project settings.