Virtual Agent Alpha Privacy Disclosures
We're building an AI-powered virtual support agent solution for Jira Service Management. For this Virtual Agent Alpha Program, we are working with select customers to help develop and train our AI models.
What does it mean to participate in this program?
In the first stage of the program, we'll analyze the content of all Jira Service Management and/or Halp support tickets and chat transcripts that have been submitted by your users to identify commonly asked questions. Our team will manually review and label the content; in return, we’ll provide you with a report of insights derived from that data.
Then, we'll combine your support tickets and chat transcripts with tickets and transcripts from other customers participating in the program to develop and train a global machine learning model. By “global,” we mean that the same model will be used for all customers, trained on data from all customers. (This is different from some other intelligent experiences you may have used, like recommended search terms, which are trained for you based on only your data.)
What if we change our mind?
We’ll rely on you as the organization admin to ensure your users understand and agree to how we’ll use the content they’ve provided in support tickets and chat transcripts for machine learning.
During the program, if you or your users decide you no longer want us to access your data, you can contact us through email@example.com. We’ll stop any analysis on your tickets, and remove your tickets from our training data set for the machine learning model. However, we may retain insights derived from your data for development of our models.
What is a machine learning model?
To learn more about how we use machine learning at Atlassian, check out https://www.atlassian.com/trust/privacy/intelligent-experiences.
What happens after the global machine learning model is developed?
In the second stage of the program, we’ll make the virtual support agent available for you to try out with your users. In return, we’ll use the chat transcripts from your use of the virtual agent to further train and improve the model. That will happen in two parts:
- First, we'll use your data to improve how the model works for you. Your data will not be used to improve how the model works for other customers in Alpha.
- Second, we’ll analyze model performance across our Alpha customers to identify common areas for improvement. Here, insights derived from your data will be combined with data from other Alpha customers to refresh the training dataset for the global model.
What if we change our mind?
Just like in the first stage, we’ll rely on you as the organization admin to ensure your users understand and agree to how we’ll use the content they’ve provided in support tickets for machine learning.
During the second stage, if you or your users decide you no longer want us to use your data, you can contact us through firstname.lastname@example.org. We’ll shut off the feature in your instance, and within 5 business days remove your data from our training data set for the machine learning model.
You represent that you have obtained the appropriate consent and/or rights in and to the data to participate in this program.