Skip to main content

Natural Language search

AskJ1 AI converts your plain text questions into JupiterOne's Query Language (J1QL), so that you can quickly find relevant information and get answers to your queries.

Using AskJ1 AI

To search for information in JupiterOne about your assets or security posture but without extensive knowledge of J1QL, simply type your question into the persistent search bar available in the upper right corner of the J1 navigation ribbon or on the J1 Hompage. Press Enter on your keyboard. AskJ1 AI will return suggested J1QL to answer the question prompt. You may edit this if something is misunderstood.

For example, you may ask "What new s3 buckets were created recently?" which will return something like FIND aws_s3_bucket WITH createdOn > date.now-7days. You can then edit the query before running the query if your definition of "recently" is the past 3 days instead of 7days.

If you are looking to continue writing with J1QL just start with FIND and the searchbar will recognize your intention to write J1QL.

note

Natural language querying is now supported on Homepage Searches as well as the Searchbar in the Navigation ribbon.

2023-09-26 15 46 37

Query Validity

AskJ1 AI baises towards query validity and may not always understand the intent of your question. Further, question prompts that are not specific or vague will not generate useful J1QL. For example, a question of "What is my security posture?" is unlikely to return useful data. A question such as "What are my open critical security vulnerabilties?" will be more likely to return a query that can answer that question.

Providing feedback

Your feedback is essential for improving the accuracy and usefulness of AskJ1 AI. If your query is slightly off or is not generated correctly, just edit your query and we will infer that as feedback to improve our model. If your query is generated correctly, just go ahead and hit enter to run the query and we will infer that a successful generation.

Data usage disclosure

AskJ1 AI uses information from JupiterOne’s app documentation, internal data models, and blogs combined with general knowledge from OpenAI’s GPT-4 model to answer questions. It does not have access to any sensitive or customer-specific data. The only details we pass to OpenAI are the user entered prompts and that information is not linked back to an account ID or other identifiable information.