The AI Chat Assistant is a guided, conversational AI assistant inside Ardoq that lets you ask questions about your reports in everyday language. It will be available throughout Ardoq in the future, but only on reports in our beta release. It summarizes a single report's content, explains what it means, and helps you gain insights faster. The assistant is secure and permission-aware as it has the same access as the user.
Availability: Open Beta. Functionality and UI may change. To get access to this feature it can be toggled on from the organization settings, provided you have Ardoq AI features enabled.
What Does it Do?
The AI Assistant lets you use an AI model to query Ardoq reports
Summarize a report to understand key insights quickly.
Example: “Give me the highlights from this Application Risk report.”Explain a report in plain language for non-technical stakeholders.
Example: “Explain what this report means for Finance.”Find action points in your data to identify problems and potential improvements hidden in your data. Example: “Give me some action points based on the data”
Why it matters: You can ask in plain English, understand reports faster, and cut through clutter with natural-language search. It is built natively into Ardoq, secure, and respects your existing permissions.
Who can use it?
All organizations that have Ardoq AI features enabled, can turn this feature on through the organization settings → features console. This has to be done by an Admin.
Any user who already has permission to view the underlying reports. Answers are permission-aware and will never expose content you cannot see.
How to access it
Open a report you want to explore.
Look for the AI assistant icon on the top right corner of the Report panel.
Type a question in plain language, for example:
“What does this report tell me about critical applications?”
“Summarize the main risks in my application portfolio.”
“What should I look at first in this report?”
Ask follow-ups. It supports multi-turn chat with message history, so you can refine or go deeper.
Give feedback on replies with thumbs up or thumbs down to help us improve quality.
No setup required, the AI assistant is native to Ardoq.
Tips for great results
Be specific. Include the scope or angle you care about, for example “for EMEA” or “past quarter.”
Use follow-ups to drill down. The assistant keeps your conversation context.
Check the source. If something looks off, then look at the underlying report to validate. The assistant summarizes what is already there.
Use small reports for best results. Smaller reports are more accessible to others, especially when exposing it to Contributor users. In addition the AI Assistant will make less mistakes when there’s less data to process.
What is included in the Beta
Single Report exploration using natural language
Summaries and interpretations of a single report’s content
Multi-turn chat
Permission-aware responses that match your existing access rights
Known limitations in the Beta
Scope is focused on reports. It is designed for report exploration first, not a general chat over all Ardoq data.
Only one report at a time. The AI Assistant cant work across multiple reports at this time.
Inaccurate calculations and math. Large language models, LLMs, are not particularly good at doing numeric calculations. This affects the report assistant feature - so please double check numeric answers to verify correctness.
Very large reports are unsupported. The assistant works best on reasonably sized reports, with up to 500 rows. Very large reports with several thousands rows, and many columns, may fail or give bad results. Future versions of the assistant will handle large reports better.
No saved chat history. Although the assistant supports multi turn chat, it does not save the history. If you reset the chat or navigate away from the report, your chat session will be lost.
Security and privacy
Responses are permission-aware, aligned to what you are allowed to see.
The assistant is built natively into Ardoq Core. No separate setup is required.
The assistant supports multi turn chat, but history is currently not saved. We plan to save chat history in future versions of this feature.
For more general information about security and privacy in Ardoq’s AI feature please see this FAQ.
Responsible use guidelines
Always verify results. AI responses are suggestions for accelerating analysis. They can be inaccurate, incomplete, or contain errors, especially when dealing with complex calculations, counts, and statistics. You must verify all numerical and critical governance data against the underlying Ardoq report before use.
Do not input any sensitive data. Avoid submitting personally identifiable information (PII), proprietary passwords, or confidential information into the chat interface. Your inputs are handled securely, but exercising caution when interacting with all AI tools is recommended.
Remember the chat is designed exclusively for report exploration and interpretation, not for general Ardoq feature usage or general knowledge questions.
Frequently asked questions
Q: Does the AI Assistant change my data?
A: No. It reads and summarizes what is already in your reports.
Q: Can it search across all workspaces and artifacts?
A: No, not yet. For the Beta, we're focusing on helping you explore and understand your reports.
Q: How do I get access?
A: First, your Ardoq organization needs to have AI features enabled. Once they are enabled, an Admin can turn on this feature in the Organization Settings → Features
How do I give feedback?
You can rate any response using the thumbs up or thumbs down icons located in the chat window. Your feedback helps improve the quality of future answers.
Thumbs Up: Use this to signal that an answer was helpful or accurate.
Thumbs Down: If an answer isn't quite right, click the thumbs down icon. You will be asked to select the issue from options like Inaccurate, Incomplete, or Error.
Providing Details: If you select Other, or if you want to provide specific context, please type your explanation in the text box and click the submit arrow.