As of 13 August 2026, AI can create a work sample test for candidates.

What to actually do

1. Hand it to a person

   The route this page recommends

   A person who owns the outcome does this end to end, worth it when the failure is dear.

2. Use a tool built for this

   Second choice

3. Do it yourself

   The distant third

   A chat interface, chat-fluent skill, and roughly 30 minutes until you can act on the result.

   How to actually do it

   1. Open the current job description and list the three to five competencies that a successful person must demonstrate in the role.
   2. Gather the assessment stage, candidate time limit, permitted tools, available information, seniority and any reasonable-adjustment requirements.
   3. Paste those details into the prompt and ask the chatbot to create the exercise, candidate instructions, rubric, example answer and scoring sheet.
   4. Compare each rubric criterion with the job description and delete any criterion that measures an unrelated qualification, personal preference or access to a particular paid tool.
   5. Give the draft to an HR colleague and a role expert to check accessibility, consistency with your recruitment process and potential discrimination or employment-law risks.
   6. Ask one existing employee to complete the exercise within the stated time, then revise unclear instructions and scoring criteria before sending the final test to candidates.

   Prompt

   Create a candidate work sample test for the role below.
   
   Role: [job title]
   Job description: [paste the current job description]
   Required competencies: [list the skills and behaviours the test must assess]
   Seniority: [level]
   Candidate context: [what candidates may reasonably know before starting]
   Time limit: [minutes]
   Permitted tools and information: [list them]
   Assessment stage: [application, first interview, final interview or other]
   
   Produce:
   1. A short explanation of what the exercise measures and why it reflects the role.
   2. Candidate-facing instructions that are clear, neutral and complete.
   3. Any fictional data, documents or scenario the candidate needs, with all facts included.
   4. A marking rubric with observable criteria and four performance levels, from insufficient to excellent.
   5. A model strong answer or worked example for the assessor, clearly labelled as an example rather than a fixed answer.
   6. A scoring sheet that two assessors could use independently.
   7. Reasonable-adjustment and accessibility considerations.
   8. A short pilot plan covering what to test with an existing employee before using it with candidates.
   
   Do not assess protected characteristics, personal background, accent, unpaid availability or access to expensive software. Do not require confidential employer information. Keep the task achievable within the stated time. Flag anything that needs a human HR or legal review. Do not invent requirements that are absent from the job description. Use plain UK English.

   Open it prefilled in ChatGPT or Claude, or copy it into Gemini, which takes no prefill link.