We’re pleased to announce the beta release of Codatum Agent, an agent that autonomously performs data analysis, creates Notebooks, and adds charts—all through natural language instructions.

Simply type requests such as “I want to visualize KPIs from this table” or “Analyze last week’s sales trend” in chat, and Codatum Agent will handle everything from SQL generation and execution to Notebook editing.

Even users who have previously spent time operating analytics tools or writing SQL can now move their analysis forward simply by having a conversation.

Two modes for different use cases

• ASK Mode

  • This mode is for when you want to inspect data or Notebook contents, or run queries to check something. It allows you to investigate and validate safely without modifying existing Notebooks.

• Agent Mode

  • This mode is for actually performing actions such as creating or editing Notebooks and adding charts.With a single instruction like “Create a graph of this metric’s trend and add it to the Notebook,” the task can be completed end to end.

Key features

• Operate with natural language

  • You don’t need to write SQL. Simply type something like “Aggregate monthly sales and turn it into a chart,” and the AI will generate and run the query, then add the chart to your Notebook.

• Automatically discovers relevant tables

  • Based on catalog information and table schema details, the agent automatically identifies the tables needed for analysis. You don’t need to manually search for the right tables yourself.

• Flexible context control

  • You can specify a Notebook or catalog as context, allowing the agent to perform analysis with reference to a particular context.

Switch models with AI Profiles based on your use case

You can switch the AI model used depending on the scale of analysis or your accuracy requirements. (Paid option)

Example use cases

• “Create a chart of sales trends by channel using last week’s order data”

  → Automatically identifies the relevant tables, generates and runs SQL, and adds the chart

• “Add a summary of monthly KPIs to this Notebook”

  → Appends content while referencing the existing Notebook

• “Investigate why the conversion rate dropped”

  → Cross-analyzes multiple tables, aggregates findings, and reports the results

• “Create a new Notebook for a sales report”

  → Automatically creates everything from the Notebook structure onward

For setup instructions and detailed usage, please refer to the support documentation.

Service site is here.

Specifications are subject to change without notice.

We hope this helps make your data analysis more effective.

Please give it a try, and we would greatly appreciate any feedback or feature requests you may have.