Build your first AI workflow: lesson 9 of 10

Testing a Workflow

A workflow you have not tested is a guess. Test it on a few real jobs, score what comes back against checks you wrote down first, fix one thing, and run it again. The loop is short, and it is what makes a workflow reliable.

Test setRun itScore itRecord itFix onethingRun itagain
Test set
Pick a few real jobs, easy, typical and tricky, and write down what a good result looks like first.

What Makes a Fair Test

  1. 1Decide what good is firstWrite the checks before the run. Criteria written after you have seen the output tend to fit it.
  2. 2Keep a hard caseInclude an easy job, a typical one and one with a complication. Keep the case that failed, and run it every time.
  3. 3Change one thing at a timeSame input, same checks, same version. If two things changed, you cannot say which one helped.
  4. 4Write the run downThe date, the tool and its version, the case, what each check found and the score. A run you did not record cannot be compared with the next one.

Pick the Better Move

Each case shows a choice in a test. Pick the better move.

  1. 1

    A run failed one check: the summary dropped a number from the report.

  2. 2

    Your test set is five easy jobs, and all five passed.

  3. 3

    When do you write down what a good result looks like?

  4. 4

    A fix for test case 2 worked.

0 of 4 answered.