Skip to main content
Version: Next

Lakehousecat AI Agent

The Lakehousecat AI Agent is a conversational assistant built into every session. It understands your data, executes analytical tasks in natural language, and — depending on your role — can manage datasources, models, users, and platform resources.

How It Works​

The Agent operates within a session. You type a request in plain language; the Agent interprets it, selects the appropriate action, and returns a result — a chart, a table, a confirmation, or a follow-up question.

The Agent is role-aware: the set of actions it can perform expands with the permissions assigned to your account. A standard user can generate charts and manage their own sessions. A builder can create datasources and models. An admin has full platform management capabilities.

Capabilities by Role​

Capability AreaUserBuilderAdmin
Generate charts and analyze data✅✅✅
Manage own sessions and prompts✅✅✅
Manage charts and dashboards—✅✅
Create and configure datasources—✅✅
Create and manage AI models—✅✅
Update semantic descriptions—✅✅
Schedule and manage jobs—✅✅
Install sample packages——✅
Manage users and groups——✅
Monitor platform instance & license——✅
Back up and restore the instance——✅
Review the audit log (who changed what)——✅

The Agent Remembers You​

The Agent maintains a personal memory for each user. When you share a preference — such as a preferred language, a chart style, or a response format — the Agent stores it and applies it automatically in all future interactions.

You do not need to repeat preferences in every session. Once set, they persist.

Examples:

  • "Always give me feedback in English." → The Agent responds in English from that point on.
  • "I prefer horizontal bar charts over vertical ones." → Applied automatically when the Agent chooses a chart type.
  • "Keep your answers short and skip the explanations." → The Agent adjusts its response style accordingly.

The Agent also learns implicitly from your interactions over time — the chart types you prefer, the level of detail you engage with, and the patterns in your requests. You are always in control: you can instruct the Agent to forget a specific preference, or to clear everything it has learned about you.

How to Interact​

The Agent responds to natural language. Phrase your request as you would ask a knowledgeable colleague:

  • "Show me a bar chart of revenue by region for the last 6 months."
  • "Create a new datasource for our production PostgreSQL database."
  • "Add user anna@example.com to the Analytics group."
  • "What is the current license state of this instance?"

The Agent infers intent from context. You do not need to know command syntax or API endpoints.

Role Guides​

For a detailed view of what each role can do with the Agent: