Watson Insights

Available as of APIC release 10.0.4

As an enhancement to the AI Test Generation feature introduced in the 2021.2.1 release of CP4I, AI generated insights now take the form of Watson Insights. Up to three test cases with an impact score of at least 70% can be recommended, as denoted by the Impact score nn% annotation on each tile.

Watson Moments Panel

Panel breakdown:

1 - Test coverage improvement

  • By analyzing the amount of occurrences covered by each of the highlighted test cases we can calculate a number for the total potential coverage improvement for your API provided by utilising all of the recommended cases.

2 - Impact score

  • An impact score should be seen as a quick "overview" of how impactful a test case would be to add. The score is calculated by comparing the overall usefulness of all cases and calculating a relative score expressed as a percentage. These values are shown in the tooltip seen in point 3.

3 - Impact score tooltip

  • The values used to calculate the impact score are shown here. Data completeness describes how complete the test case is based on the values that are pre-populated for you. Occurrences covered describes how many traces from this set of results were covered by this single test case. The final Impact Score is calculated by taking these two values, comparing them with the other generated insights, normalising the data and combining them to calculate a geometric mean between the two values which creates our final score*.

*This value may change in future updates as more data is used to further enhance AI Test Generation.