FAQ for Copilot in Power Automate Process Mining
What is Copilot in Power Automate Process Mining?
Copilot in Process Mining enables users to streamline data ingestion and provides process insights through quick and easy natural language expression. With Copilot in ingestion, identify your process during data ingestion and auto map your data to the required data schema. Copilot can also surface insights on your process and recommend solutions through natural language in Power Automate process mining.
What are the systems capabilities?
The system is a copilot embedded in process mining ingestion and the desktop application. It helps you get to actionable insights of your process faster. The system can:
- Discover your selected process in Bring Your Own Azure Data Lake.
- Give automapping recommendations to required data schema.
- Surface top insights in your process.
- Offer recommendations on automation.
- Answer questions on your process data.
- Answer general questions about processes.
- Filter out questions not related to the analyzed process or processes in general.
What is the system's intended use?
The systems intended use is to help you ingest your data into process mining and help you surface actionable insights easily and quickly.
How was Copilot in Process Mining evaluated? What metrics are used to measure performance?
We have a robust set of metrics we're tracking to measure the performance of the model and resulting customer experience. We track Copilot's Service Level Agreement (SLA) to make sure it's always available to you. We track the telemetry of thumbs up and thumbs down gestures present in the AI Copilot for each AI output that you can submit feedback for.
You can provide feedback for Copilot when the results are biased or inappropriate. We track this feedback to ensure that Copilot is compliant, appropriate, and bias-free.
What are the limitations of Copilot? How can users minimize the impact of the Copilot limitations when using the system?
Copilot can support ingestion only in a process created through Azure Data Lake as a source. It supports only analytics driven through the Power Automate Process Mining desktop application. It isn't equipped with transforming or preparing any ETL work on behalf of the users. You're made aware of these limitations throughout the experience with warnings that AI outputs must be validated. You'll also only be surfaced copilot prompts and experiences in the sections of the products without limitations.
There's also an option to disable the Copilot within the admin center, and the user always has the option to close the Copilot and not engage with the feature.
What operational factors and setting allow for effective and responsible use of the system?
Copilot is only available in the Azure Data Lake ingestion experience and within the Power Automate Process Mining desktop app. The user can always close Copilot and not elect to interact with Copilot.