AI agents help analyze information and generate insights, but they may occasionally produce inaccurate or incomplete results. Outputs are based on the quality, completeness, and availability of underlying data and sources.

For the most reliable results, it is important to maintain well-structured and verified data sources within your organization — for example by organizing datasets, references, and documentation in tools such as the DataHub and dataset libraries.

Always review and validate important outputs before using them in reports or decision-making.

Key Principles

• AI results may contain inaccuracies

• Outputs depend on the quality and availability of data and sources

• Well-structured datasets improve the reliability of AI insights

• Always review important outputs before using them

Keeping Your Sources Relevant Over Time

To avoid building up outdated or legacy sources, we recommend maintaining a structured approach to managing your datasets and references.

Best practices include:

1. Regularly review key datasets
Schedule periodic reviews of important datasets to ensure sources, assumptions, and references are still valid.

2. Document sources clearly
Always attach documentation, links, and context to your datasets so future users understand where the information comes from.

3. Replace outdated information
When newer or more reliable data becomes available, update datasets rather than creating parallel versions.

4. Maintain a clean dataset library
Archive or remove unused datasets to keep your dataset library clear and relevant.

5. Assign ownership
Ensure datasets have responsible owners who periodically review and update the information.

Good data leads to better insights.
Well-maintained sources and structured datasets significantly improve the reliability and usefulness of AI-generated results.