Best practices

Assisted Task Mining is designed to help you capture additional details and information about a known task to get a comprehensive picture of different ways the task is completed.

Understanding the why

When starting to use Assisted Task Mining, a best practice is to always explain to your end users your rationale for creating an ATM project and what you hope to achieve by recording the identified task. Assisted Task Mining records everything on the recording user’s desktop so setting up the task environment before recording is vital to limiting noisy data.

Key callouts include:

• The recording will only start with the end users' consent.

• Only applications that have been allowed will be recorded.

• You can configure Personal Identifiable Information masking (PII).

• User anonymization is always an option. The goal of recording this data is to better understand the inefficiencies in the task to further optimize the team/task and drive actionability.

Project setup

Navigate to your Task Mining service and create an Assisted Task Mining Project. It's best to create a project per team/department. During project creation, you should provide a unique and relevant title and description. For best results, it's best that users are experienced and familiar with the day-to-day job.

Each user will receive an email notifying them that they've been added to the project and that they can access the project in UiPath® Automation Cloud^TM to record or review traces.

Be specific with the recording users that you invite and indicate the exact known task with relevant applications that you'd like them to record. This helps in limiting noisy data when reviewing or merging your traces.

Recording tasks

Each project can have up to 100 total traces. Traces can have up to 500 actions.

• The recording will only start once the user has authenticated themselves on the recorder and clicked Start.

• All applications on the desktop that are visible could be recorded so users should be aware and only have the applications they’d like to capture available.

• PII masking is turned on by default and will mask all screenshots.

Tips for recording:

• Finish the task that you started. This minimizes interpretation/breaks during the recording.

• Limit task switching, if possible, while recording, and minimize the usage of non-supported apps.

• Pause recording if other things come up during the completion of a task.

Analyzing your traces

Once traces have been recorded and uploaded to the project, you or anyone with a Business Analyst or Admin role within the project can review them. When reviewing the trace, it’s important to make sure that the trace you recorded includes all necessary parts of the identified task. Note that PII masking will be applied by default so the images you review in your trace will be masked.

Merging your traces

Once all the traces for the project have been collected and uploaded, you can merge traces together to get a comprehensive view. There are two types of merges: Auto merge and guided merge.

• Auto merge: Leverages AI to find the similarities between steps and their actions of each trace for you. It'll then create a comprehensive view of the traces together and automatically create decision points where actions are different. Up to 25 traces can be auto-merged together, each trace having up to 500 actions collected in it.

• Guided merge: Enables users to merge multiple traces together on their own by appending and merging individual steps to create a comprehensive picture of the trace. It's recommended to have a Subject Matter Expert who understands the business context of the task at hand to complete this for the team.

When reviewing your trace, note that any additions or deletions to the step will make your trace ineligible for auto merge. You'll still be able to later complete a guided merge, but you can only change the trace name for auto merge.

Driving actionability and process optimization

Once the team has a comprehensive view of the task at hand, it's recommended that a Business Analyst review the merged trace and fill in any potential details that'd be beneficial to the larger group. For example, the Business Analyst might change step titles to be more relevant for business users as well as provide descriptions that include context of the action being completed.

Once reviewed and edited, the merged trace can be exported to a Process description document (.docx), an image, or a studio (.xaml) file to accelerate your discovery journey.