Skip to main content
In a private (Docker) deployment, you must deploy aisearch before this feature works — see Deploying AI Search.

​
Overview

This feature lets you ask questions using abbreviations, such as “Alibaba” or “Mate 40”. The system automatically converts them into the full company or product names and generates the correct query, so you retrieve data accurately. It suits scenarios where the database stores full names but users habitually ask with abbreviations.

​
How to use it

On the Data Overview tab of the data source detail page, select a table, then use the AI Index column in the field table to control fields individually:
  • Default settings: the system automatically enables AI Index for some fields. By default, these are short text of string type — not long text, dates, or numbers. These fields help the AI match user questions more accurately.
  • Manual configuration: click a field’s toggle to enable or disable its AI Index. When enabled, a status icon appears next to the toggle.

​
Viewing AI Index details

Click the info icon in the AI Index column to view the field’s index status and statistics, including:
  • Total records
  • Indexed value count
  • Distinct count
The system uses these statistics to decide whether a field meets the AI Index requirements.

​
Why a toggle?

Enabling too many indexes can affect performance, since each index consumes computing resources. A large amount of irrelevant index information also becomes noise that interferes with AskTable’s analysis. The toggle lets you enable or disable indexing per field based on your needs.
  • Enable AI Index: the AI understands and matches your questions more intelligently, improving query accuracy.
  • Disable AI Index: improves performance, but may sacrifice some accuracy.

​
Summary

  • Enabled by default: suitable for most scenarios; the system automatically enables AI Index for certain fields.
  • User configuration: you can manually enable or disable AI Index per field for flexible adjustment.
  • Performance optimization: too many indexes may affect performance, so configure as needed.
By configuring AI Index, you can improve the flexibility and accuracy of AI queries while optimizing performance for your actual needs.

​
Notes

  • Enabling AI Index may increase computational load, so enable it only when necessary.
  • The default-enabled fields are already optimized and suit most scenarios.