Learning AI · Practical guide
Test what happens when a form cannot be completed.
Use a fictional supply-request form with name, quantity and an optional description. Check both field rules and whether someone can correct an error without starting again.

01Define acceptable values
For this exercise, name is required, quantity allows integers from 1 to 10, and description may be empty. Write these rules before testing so unexpected behaviour does not accidentally become the requirement.
Try an empty quantity, 0, 1, 10, 11 and 1.5. Under this contract, boundary values 1 and 10 are accepted; the others require the intended invalid-value behaviour.
02Check whether an error helps correction
Submit with an empty name. Expect a message identifying the field and explaining what is missing. “An error occurred” does not distinguish incomplete data from an unavailable service.
Correct the name. Check that other fields retain their values and that the message clears or updates on validation. Navigate through the fields with a keyboard too.
03Distinguish interface and server validation
MDN explains that browser validation does not replace server validation. In a real application, ask the builder where rules are enforced when data arrives.
In an authorised test environment, the team can submit an out-of-range quantity directly to the entry point and check that it is rejected there too. A disabled button alone does not establish this protection.
Reference [1]: MDN Web Docs
04Rehearse a failed save
Arrange a simulated save failure with the team in testing. The form should not claim that a request was created without confirmation. Define the message and what the person can do with the data already entered.
Record field, input, expected outcome and observed outcome. After a fix, repeat boundary values and a complete valid case. Keep results with acceptance criteria.
Sources and further reading
These references expand on the concepts indicated. The examples and exercises are original editorial material.
[1] MDN Web Docs
Client-side form validation ↗
Browser and server validation.
Back to the related section
Frequently asked questions
Should I test using real customers?
Use fictional data in a test environment. Agreeing how to simulate the failure avoids creating real requests or changing other people’s information.
Keep reading
AI for business
Acceptance criteria: how to check an application ↗
Replace vague requirements with checkable examples. Define inputs, actions and outcomes using a request-approval exercise.
Learning AI
How to check an AI assistant’s answers before using it ↗
Build a simple test with questions, evidence and acceptance criteria. Use a worked exercise to find unsupported answers before relying on an assistant.
Learning AI
Turn meeting notes into tasks with AI: a checkable exercise ↗
Practise extracting commitments, owners and deadlines without inventing them. Includes fictional notes, a reusable instruction and a reference answer.