How Do You Prove AI Skills to an Employer?

A certificate and a work sample answer different questions

A certificate can show that you completed training or met an assessment standard. A work sample lets an employer inspect how you approach a task.

If your CV says you use AI for research, analysis, or writing, prepare an example that makes that claim concrete. What were you trying to achieve? Where did AI help? Which parts needed checking? What did you decide?

The aim is to make your capability easier to assess.

Choose a sample that matches the role

Start with a task the employer is likely to recognise. A small, well-explained sample can be more useful than a large collection of unrelated outputs.

For example:

• Marketing: a content-recovery recommendation that connects search data to a decision about a declining page.
• Product management: a customer-feedback brief that separates recurring problems from isolated requests and explains the prioritisation.
• Operations: a process-improvement proposal that identifies a bottleneck, compares options, and records the constraints.
• Software: a change that includes the problem, implementation, tests, and an explanation of how AI-generated suggestions were reviewed.

Use work you have permission to share. If you create a sample using public or synthetic data, say so. It can still demonstrate your approach without implying that a real organisation adopted it.

Make the work easy to inspect

Package the sample with a short explanation covering five things:

1. The task: what needed to be done, for whom, and under which constraints?
2. The deliverable: what can the employer actually inspect?
3. The AI assistance: what did AI help draft, organise, analyse, or build?
4. Your checks and decisions: what did you verify, reject, correct, or choose?
5. The result and limits: what happened, and what has not yet been tested?

Include the relevant evidence, such as source material, calculations, tests, or a short comparison of the initial and final versions. Select the parts that explain the work rather than attaching an entire chat history.

Show a decision, not just a polished output

A finished document may not reveal whether you noticed a missing constraint or corrected an unreliable answer.

Pick one meaningful decision and explain it. For example:

This gives the employer something specific to discuss: your checks, your reasoning, and the consequences of the decision.

Be ready to explain an alternative you rejected and what additional evidence might change your recommendation.

Keep the outcome accurate

Separate producing a deliverable from achieving a business result.

If you recommended a change, say that. If the change was implemented, explain what happened. If you claim that AI saved time, include the time spent checking and correcting its output.

Avoid presenting an estimated saving as a measured result or a practice exercise as client work. Clear limits make the sample easier to evaluate.

Use the sample in an application or interview

Add a descriptive link to your CV or application, such as "Customer-feedback analysis with prioritisation rationale."

Include one sentence explaining its relevance to the role. In an interview, walk through the task, one important decision, and the evidence supporting the result.

Ask yourself before sharing: can someone understand what this demonstrates without needing a long explanation from me?

Build evidence through practice

Altiv.AI Plays provide structured practice and updated learning content around concrete tasks. You work towards a useful deliverable while learning where AI can make the work faster, improve its quality, or enable a different approach.

If you are unsure which tasks to start with, use the AI Readiness Score to explore tasks classified as automated, augmented, or human-only. Use that assessment to choose what to practise, then build a work sample that shows what you can do.