Train the AI

Train the AI allows you to fine-tune your machine learning models.

Functionize uses machine learning to keep your tests up to date as your application changes over time. This means that your tests are dynamically changing to interact with your UI elements even as they change.

Sometimes, the test self-heals to interact with the wrong element. In these cases, you can help train the AI by reporting the error to our data science team. Submit issues with the models directly from the action that uncovered the error so that we can include additional training data and verify that the machine learning model is adapted. 

Train the AI

  1. Open a test in your Project. Open Slider View via the View Results button or the View button on any Action.
  2. Select the Action you wish to report. Open the Action Menu and select Train the AI.
    Train1.png
  3. Select what type of error occurred. Optionally, you can also provide comments and allow the Functionize team to rerun your test for verification.
    Train2.png
  4. Click Submit
    Train3.png

Report via Orchestration Result Details

You may also submit Actions for review when you access your execution details from the Orchestration Results page.

  1. Open the desired Orchestration Results.Train4.png
  2. Click the details button for any test to open that test's execution.
    Train5.png
  3. A new tab will open with a read-only Test Detail page for this execution. Click View Results for a step-by-step read-only review in Slider view.Train6.png
  4. Select the Action you wish to report. Open the Action Menu and select Train the AI. Follow the submission procedure outlined above.
    Train7.png