Artefacts
During complex conversations or background workflows, agents often generate significant amounts of valuable information, such as comprehensive status reports, diagnostic summaries, or structured charts.
To ensure this information is not lost in a conversation, DataQI allows agents to generate and store Artefacts.
What is an artefact?
Section titled “What is an artefact?”An Artefact is a unique, retrievable entity produced by an agent.
Unlike a simple chat message, an artefact is stored in the DataQI database. This means it can be queried and retrieved by the platform entirely separately from the conversation or workflow that initially created it. Crucially, all artefacts are securely bound to the session and user that created them, ensuring your generated data remains strictly private.
DataQI currently supports generating the following types of structured artefacts:
| Type | Description | Export format |
|---|---|---|
| Datasets | Dynamic data tables (including search result sets). | .csv |
| Charts | Visual data representations, including Donut, Bar, Line, and Mermaid relationship graphs. | N/A |
| Documents | Standalone markdown reports and summaries. | .docx |
Artefact generation
Section titled “Artefact generation”Agents can generate artefacts in two ways:
- Chat sessions: An agent might generate a “Weekly Marketing Summary” document artefact and attach it to your ongoing conversation. While you can view the artefact immediately, its independent persistence means that DataQI agents can retrieve and reference it later in subsequent workflows without losing context.
- Background workflows: When agents execute complex tasks (such as performing root-cause analysis on an infrastructure incident), they generate artefacts to summarise their findings. A strong example of this is a relationship graph (Mermaid chart) illustrating network dependencies or system states. These artefacts are automatically attached to the specific workflow run and surfaced in the Agent feed.
Persisting agent value
Section titled “Persisting agent value”The primary goal of Artefacts is to capture and save the value generated by AI agents. By saving these reports, charts, and summaries, DataQI ensures that critical insights are preserved — allowing agents to securely build upon previous work and context long after it was created.