How to List AI Skills on Your CV (ChatGPT, Copilot, Prompting) Without Sounding Vague

«Proficient in AI» is the «proficient in Microsoft Office» of 2026: technically true of everyone, convincing to no one. Yet employers do want to know whether you can use these tools well, and job ads increasingly say so. The difference between a line that helps and a line that hurts is whether it describes a result. Here is how to write the helpful kind, by profession.

Name the tool, the task and the outcome

The formula is the same as for any skill: tool + what you did with it + what changed. Not «Experienced with ChatGPT» but «Used ChatGPT and Claude to draft first versions of client proposals, cutting turnaround from three days to one while keeping final review and compliance checks manual.» The last clause matters: it shows judgement about where the tool stops.

Where it goes on the CV

• In the skills section, as named tools under a sub-heading («AI and automation: ChatGPT, Microsoft Copilot, Claude, Midjourney, Zapier»), only tools you have actually used for work.
• In the experience bullets, as results, which is where it carries weight.
• In the summary, only if the role asks for it («… and two years of building AI-assisted reporting workflows»).
• Never as a job title you did not hold («Prompt Engineer») unless it was.

Examples by profession

• Marketing: «Built a Copilot-based workflow for first-draft campaign copy in three languages; A/B tests showed equal conversion at 40 % lower production time.»
• Finance: «Automated monthly variance commentary with Excel Copilot and reviewed every figure against the ledger before distribution.»
• Customer service: «Trained the team on an AI reply-suggestion tool; average handling time fell 18 % with CSAT unchanged.»
• Software: «Use GitHub Copilot daily; introduced a review checklist for AI-generated code that caught two security issues before release.»
• HR: «Drafted job descriptions with ChatGPT and screened them for biased language with a checklist; time to post fell from five days to two.»
• Teaching: «Designed AI-use guidelines for a class of 28 and built rubrics that assess reasoning rather than output.»

A recruiter does not want to know that you used AI. They want to know that you knew what to do with the answer.

Phrases to avoid

• «AI expert», «AI enthusiast», «passionate about AI» without a single result.
• «Prompt engineering» as a skill when it means asking a chatbot questions.
• Listing twelve tools you tried once.
• Claiming you «built an AI» when you used one.

Certificates: do they help?

Short vendor courses (Microsoft, Google, Coursera AI fundamentals) are worth one line under Certifications if the job ad asks for AI familiarity; they do not replace a result in your experience section. Formal credentials (a data-science degree, an ML engineering certification) belong in Education or Certifications with the year.

Country notes

In Germany and France, employers increasingly ask about AI use in the context of data protection; one phrase about handling confidential data correctly («no client data entered into public tools») reassures. In the Gulf, government and large employers list AI literacy explicitly under national digital programmes; use the same words. In Japan and China, name the tools in use locally (Copilot and domestic models) rather than only Western ones.

Say it once, say it with proof

CV Analyz checks your skills section against the job ad, flags vague claims and rewrites tool mentions as results, using what you actually did.