Help Centre › Building Processes
A run that finishes is not always a run that did the right thing. A process can finish cleanly and still post the wrong amount. Checks let a process say what "correct" means, so a wrong result shows up as a failure.
1. Add check steps to the process
Add a check where the result can be seen:
- Check that a condition holds checks one or more conditions on values, such as "the total is above zero", and fails the run if one does not hold.
- Verify page conditions checks conditions on the web page the bot is on.
A check that fails stops the run like any other failed step, unless the step is set to continue on fail. Then the run carries on and the failed check is still recorded.
2. See the checks of a run
Open the process, choose the Runs tab and pick a run. If the run passed through any check steps, its Checks section lists each one, failures first, with the value that failed.
A run with no check steps shows no Checks section at all. A run can be green at the top and still have a failed check inside, when the check step was set to continue on fail. That is the point of the section.
3. Run saved cases from the bot
The bot app has a Checks page in its menu. It runs a process against cases: known inputs you save once and run again after every change, to see that the process still gives the right result.
- Sign in on the bot and open Checks.
- Pick the Process.
- Click Add case. Give it a Name, an Environment (optional), and its Inputs, one per line as name = value, for example
invoiceNumber = INV-1001. - Click Save. Add as many cases as you need.
- Click Run checks.
Each case runs the process for real, one after another, and shows its result:
| Result | Meaning |
|---|---|
| Passed | The run finished and every check held. |
| Failed | A check did not hold, or the run itself failed, was stopped or was refused. |
| No checks | The run finished but checked nothing. This is not a pass: add check steps to the process. |
| Suspended | The run is waiting for a person, such as an approval, so it has no result yet. |
| Error | The case itself could not finish, for example it ran past its 30-minute limit. This says nothing about the process. |
The tiles at the top count the Cases, how many Passed and Failed, and the Duration. Use Edit and Remove to change the cases.
What a check run leaves behind
A check run drives the real applications, because a run that does not touch them has tested nothing. What it does not do is leave a trace: it uses no credits, writes no run history, sends no log or screenshots to Automize, and does not hold a bot licence while it runs. So a long list of cases cannot lock your other bots out.
Because it does drive the real applications, point your cases at a test environment where you can, using the case's Environment.
Cases stay on the machine
A case's inputs are often real data, such as invoice numbers and account references. So cases are stored on the machine where you saved them and are never uploaded. A case saved on one bot is not on another.
Results are not kept either. The list reads Not run until you run it again, so an old green tick never sits next to a process that has changed since.