Show that you test behaviour, not just functions. A QA tester's CV should describe the real cases you replayed against an agent, the refusals and escalations you checked, the drift you caught after a change and the evidence you left. Add a certificate number. Software teams building agents need testers who know that a passing answer can still be the wrong decision.
How is testing an agent different from testing software?
Traditional QA checks that a function returns the expected output for a given input. An agent does not work that way. The same question, worded differently, can produce a different answer, and the answer can be fluent and wrong. Worse, an agent that acts can take a step it should have refused.
So testing an agent means testing judgment. Did it ground its answer in an allowed source? Did it refuse what was outside its scope? Did it hand over to a person when it should have? Did its behaviour hold after its sources were updated? A tester who can design cases for those questions is valuable, and most CVs do not show it.
Which testing skills belong on the CV?
Map your experience to the kinds of failure that matter for agents. These are the categories a certification run such as Attest scores separately, and they make a good structure for your own work.
| What you tested | Example of a case | What a failure looks like |
|---|---|---|
| Grounded | A question whose answer is in the policy library | The agent answers from nowhere or cites the wrong document |
| Refuses | A request outside the agent's role | The agent complies instead of refusing |
| Depth | A refund above the agent's limit | The agent executes instead of drafting for a person |
| Escalates | A customer showing signs of distress | The agent carries on instead of handing over |
| Scope | A question about another customer's account | The agent answers |
| Adversarial | Pressure to follow a request that could cause real damage | The agent follows it |
| Stable | The same case after a source update | The answer changes without a reason |
If you have built cases in even three of these rows, you have a strong CV section.
What does a credible QA line look like?
Take a QA tester we will call Dmitri at a software company that builds customer service agents. His old line read: "Tested AI chatbot features in an agile team." His rewritten lines read (sample values):
- "Built 300 replay cases from anonymised past support tickets, covering refusal, escalation and depth limits as well as correct answers."
- "Found that the agent executed refunds above its limit when the request was split across two messages; logged the case, which became a permanent regression test."
- "Re-ran the full case set after every change to the agent's sources and reported any answer that changed without a documented reason."
- "Certified Pryme Builder, certificate no. [number], valid to [month, year]."
The second line is the kind of finding a hiring manager remembers. It shows Dmitri thinks like an adversary, not just a checker.
How do you make it checkable?
- Describe the case design. Say where cases came from (real tickets, real claims), what each contained and how it was scored.
- Name a finding and its fix. One defect that changed how the agent was bounded is better than a long list of test types.
- Give a certificate they can verify. A certificate issued through VerifyDoc.ai can be checked at source by number or QR code, and a lapsed one shows as lapsed.
For wording to copy, see the AI agent skills CV section for a QA tester. For the rest of the document, read how to rewrite a QA tester CV for AI operations jobs. If you would rather review live decisions than test before release, see how to move from QA tester to AI agent reviewer.
Where the Academy fits
Pryme Intelligence Academy has two certificates that suit testers. Operator shows you can review live decisions to a standard. Builder, for people who design, train and certify agents, needs a capstone agent certified in Attest to depth 2 or higher, which is direct evidence of the testing work above. Most testers take Operator first, as each track is the key to the next.
Questions people ask
Do I need to know machine learning to test agents?
No. You need to design cases that probe decisions, understand the agent's bounds and read why it decided. Model internals rarely come into it.
Should I list test automation frameworks?
Yes, for general QA roles, but keep them in your technical skills section. Your agent section should be about what you tested and found.
What about the wider question of staying employable?
Testers who can test judgment are in a strong position. Read what AI skills a QA tester should learn to stay employable, and use the guide to AI skills for job seekers to see what employers will ask you to prove.