Autogenerate tests

​

Test autogeneration at sign-up

When you sign up to Octomind, our AI Agent is going on a discovery journey of your site. It takes a screenshot and tries to come up with a few relevant tests to start with automatically.

​

Cookie banner and login test

First, it’s checking if there is a cookie banner and a required login that need to be tested. “Required login” is relevant for SaaS apps that unfold most of their functionality to logged in users.

If the AI Agent finds a required login, it will ask you for test credentials. It will use them to generate and run a login test. It will be added as dependency for other tests. So will be the cookie banner test if one is found.

We can handle 2-factor authentification (2FA) and captcha in login flows.

After these tests are autogenerated, we will validate them to see whether they work. If passed, they will be set as active meaning they will run when you execute your test suite.

​

First batch of tests is generated automatically

Here comes the really cool part. Once we finished searching for a potential login and cookie banner test, we start generating 3 test cases for you automatically. You can follow the generation progress in the Agent progress bar:

It’s possible that the Agent fails to generate a test (here are some reasons why) or needs help with something - the test is highlighed in yellow. This is a step by step guide on how you’ll help the Agent by nudging it in the right direction.

​

Spawing existing test cases

Similar to the initial AI test generation during project setup you can use the generate more from here feature and create a new test from here to expand your test case suite.

Go into the test case view and click on the plus icon. Two options open:

With the first - generate more from here - our AI Agent will discover and auto-generate up to 3 new tests “following” the test case you launched it from. The original test will be added as a dependency automatically.

The second option - create new test from here - allows you to prompt the Agent with a specific test flow instruction to create a test “following” the test you launched it from.

You can follow the progress of these newly created tests in the AI Agent progress bar.

At the end of the agent run, the Agent should have autogenerated test steps and validated them for each test case. For those that the validation succeeded, it turned the test ON - into active mode.

Help the Agent when it couldn’t quite nail the autogeneration process - yellow alert highlights a failed step generation. See how you help the Agent to generate a test case successfully.