Datasets

In DataQI, a Dataset is a live, organised list of results generated when an AI agent searches your data. It is not a static folder of files, but rather a snapshot of data that matches a specific intent.

Dynamic generation

When you ask an agent to find information, it searches across your connected systems and compiles the results into a specific Dataset. This Dataset serves as the single source of truth for that specific query.

By organising search results into a Dataset, DataQI can easily manage hundreds or thousands of records without confusing the AI.

Tool chaining and refinement

Datasets are easy to work with. Because they are structured, agents can perform multiple steps on them in the background before presenting the final result to you.

For example, an agent might run a broad search to generate an initial Dataset of “All 2024 Sales Contracts”. Rather than returning that massive list immediately, the agent can use other tools to filter that existing Dataset—refining it down to “Only contracts exceeding £100k”.

This ability to chain tools and refine Datasets in the background allows DataQI to answer complex, multi-step queries accurately and efficiently.

Exporting your data

Because Datasets are structured, they are highly exportable. When a Dataset is presented to you in the chat interface (typically rendered as an interactive table), you have the ability to instantly export the entire underlying Dataset to a .csv file.

This ensures that once DataQI has done the heavy lifting of finding and filtering the data, you can easily download it into tools like Excel for further analysis.