Table of Contents

An AI agent connected to your SeaTable base can do far more than simple database queries. In this article, we show you concrete use cases — from simple queries to more complex tasks.

All examples work with any MCP-compatible AI assistant (Claude Desktop, Claude Code, Cursor, or other compatible clients).

The most common use case: you ask a question and the agent returns the matching data from your base.

  • “Show me all open tasks that are due by the end of this week.”
  • “Which customers from Berlin placed an order last month?”
  • “List all projects with the status ‘In Progress’.”
  • “Which entries in the Invoices table have no payment received?”

The agent automatically identifies the right tables and columns — even if you don’t type the names exactly.

The agent can evaluate your data and create summaries that would otherwise require building reports manually.

  • “What was the total revenue last quarter? Broken down by sales representative.”
  • “How many new contacts were created per month this year?”
  • “Which employee has the most open tasks?”
  • “Show me an overview of project budgets — planned vs. actual.”

Particularly useful when you’ve inherited a base from someone else or need an overview.

  • “Describe the structure of my base: what tables exist and how are they related?”
  • “What column types does the Projects table have?”
  • “Are there links between the Contacts table and the Projects table?”
  • “Which columns in the Tasks table are required fields?”

With a read-write token, the agent can create new entries in your base. The AI assistant asks for confirmation before every change.

  • “Create a new contact: Name ‘Müller GmbH’, City ‘Hamburg’, Status ‘New’.”
  • “Add a new task in the Tasks table: Title ‘Create proposal’, Responsible ‘Lisa’, Due ‘2025-03-15’.”
  • “Create an entry in the Invitations table for every customer from Berlin with the note ‘Invitation to trade fair’.”

The agent can modify existing entries — individually or in groups.

  • “Set the status of all overdue tasks to ‘Escalated’.”
  • “Change the email address of Müller GmbH to info@mueller-gmbh.de .”
  • “Update all projects with the status ‘Completed’ that are older than one year to ‘Archived’.”

Use the agent to check data quality without going through every row yourself.

  • “Are there contacts without an email address?”
  • “Which tasks have no due date?”
  • “Are there duplicate entries in the Customers table based on the company name?”
  • “Which projects have a start date that is after the end date?”

The agent can identify relationships and draw conclusions that go beyond simple queries.

  • “Which customers haven’t placed an order in the last three months?”
  • “Compare revenue from Q1 and Q2 — which products grew, which declined?”
  • “Which employees have tasks in more than three projects at the same time?”

Start with read queries. Before having the agent modify data, experiment with queries and analyses. This helps you learn how the agent interprets your base.

Use context. The agent remembers the conversation. You can build on previous answers: “Show me the details for the first entry” or “Filter this list by status ‘Open’”.

Combine steps. Complex tasks are best broken down into individual questions. Ask first, check the result, then give the next instruction.