Collecting Screenshots with the Monitoring Agent

The monitoring agent captures screenshots during specific user actions by running research campaigns. These screenshots can be stored, processed, and analyzed. They help speed up Task Mining project setup and improve the accuracy of activity labeling. Screenshot collection is available only with the extended monitoring license.

The monitoring agent captures a screenshot when a user interacts with on-screen forms. Screenshots are taken for the following actions:

• Clicking a button, link, or other element — activity code 52
• Editing a field — activity code 53
• Copying text — activity code 60
• Pasting text — activity code 61
• Selecting a value from a combo box, or clicking a radio button or checkbox — activity code 62
• Selecting a value from a list or table — activity code 65

The agent highlights the active UI element:

• With a rectangular frame if the element has coordinates larger than 5×5 pixels
• With a circle centered on the click location if coordinates are unavailable or too small

The screenshot ID is recorded in the screenshot_id column of the «monitoring_activity» table. Screenshots are stored in the built-in Operavix file database and are removed after the study ends. The average screenshot size is approximately 50 KB.

The monitoring agent sends screenshots together with activity data to the Operavix server. Each archive includes 10 minutes of user activity. The archive size depends on the number of captured events, which varies by business process. You can estimate the size of a daily archive using the formula: (event count × average screenshot size) × 6 × 8, where:

• umber of events × average screenshot size — 10 minutes of activity
• number of events × average screenshot size × 6 — one hour of activity
• 8 — eight working hours

Screenshot Capture Limitations

There are some limitations in screenshot capturing:

• Click events:
  • Keyboard shortcuts are logged, but screenshots are not created.
  • Clicking on context menu items closes the menu before the screenshot is taken, so only the highlighted screen area is captured.
• Field editing: if the element lacks coordinates, no screenshot is taken. Mouse coordinates are ignored for these actions.
• Value selection: screenshots are not taken for combo box items without coordinates.
• Copying text: screenshots are not created if the copied element isn’t a parameter element or lacks coordinates.
• Pasting text: same restrictions as copying apply.
• Copy and Paste events are logged via clipboard changes, which might be captured only after the next UI action, making screenshot capturing unreliable. If triggered via context menus, screenshot inheritance from the previous event is used.
• Screenshots are not captured in the main windows of Microsoft Excel or Word due to disabled window focus detection.

Configuring Screenshot Capture

To configure screenshot capture, navigate to Research in the system settings menu.

To start a new research campaign, click + Add in the top-right corner, name your campaign, and confirm with Add.

Once created, the research settings page opens. You can configure:

• Name — unique identifier for the campaign
• Start date — when screenshot capturing begins
• End date — when screenshot capturing ends
• Blur — enable or disable screenshot blurring
• Status — current status:
  • Pending — not yet started, needs configuration
  • Active — currently capturing screenshots; can be paused until the scheduled end
  • Completed — capturing ended either due to end date or manual stop

If blurring is enabled, everything except the active element is blurred. If the element is inside a circle or should be hashed, the element is also blurred.

Blurring level is set in monitoring configuration file. Default is 50. Valid range: 1–100. Set to 0 to disable blurring.

Adding Users to a Research Campaign:

1. Go to the Users tab
2. Click Add in the top-right corner
3. Select users and/or departments. To select individuals from a department, expand it and check specific users
4. Click Apply

To launch the research campaign, click Launch on the settings page.

Once active, screenshots are collected based on the defined parameters. User list and settings are locked while the campaign is running. The Launch button changes to Complete.

After a campaign ends, the agent stops collecting screenshots. You can modify all settings (except start date and user list) for manually stopped campaigns. The Complete button switches to Launch. Campaigns that finish by schedule cannot be edited.

The research ID linked to a screenshot is stored in the «monitoring_activity» table under the research_id column.