Anyone has experience with Computer vision activities instead of modern activities ? Computer vision seems to be better option, when Applications are changing frequently
Hey @keeswirken
No need to choose between the two now 
Since 23.10, we have CV very healthily integrated in Unified Target: when selectors are available, CV acts as a run-time fallback (also outputting a selector “flat tyre” warning if needed); if no selectors are available (e.g.: Citrix automation with no remote run-time installed), CV acts as the primary targeting method. Here’s a very helpful doc page that further explains the very cool targeting mechanism in place.
We still haven’t ported all the functionality from the dedicated CV activities in Unified Target, but we’re working on it. Here’s a summary of the functionality gaps:
- Only in CV Activities [WIP to be ported to UT as well]
- Run-time Auto-Scrolling
- Cell Targeting
- OCR & Area Selection Text Extraction
- Extracting scrollable text in CV Get Text
- Only in Unified Target
- Object Repository support
- Table Extraction
- Extract tables that are spanning across multiple pages
- Extraction Preview
- Get Attribute for checkboxes & radio-buttons [checked/unchecked]
Selectors are great for run-time execution speed and low-level control (you get to decide the attributes configuration, strict/fuzzy, etc.), while CV is there to boost resiliency, acting as a fallback.
Hope this helps, let us know what you think. 