Add AI skills to your CV or resume by describing a task you actually completed, the tool or method you used, and how you checked the result. Put those details in an experience or project bullet; a list of chatbot names on its own says little about what you can do.
If you are changing roles or applying for an internal opportunity, start with a task you already understand: a project update, a customer reply or a research brief. Explain your contribution without sharing confidential work. If your experience comes from independent practice, label it accurately.
Which AI skills belong on your CV?
Start with the role, then choose evidence that fits it. An operations coordinator and a marketing manager need different examples, even if both use the same AI tool.
| Skill | A task you could demonstrate | Evidence to keep |
|---|---|---|
| Writing and refining prompts | Turn a vague request into a brief with constraints | Your first prompt, revised prompt and explanation |
| Reviewing AI output | Find unsupported statements in a draft | A checked draft with sources or corrections |
| AI-assisted research | Build a short briefing from public documents | Source links, dates and notes on missing evidence |
| Drafting and editing | Adapt a message for a particular audience | The starting material and your edited final version |
| Spreadsheet assistance | Ask for a formula, then test it on sample data | The formula, expected results and error cases |
| Workflow design | Separate a repeatable drafting step from a human decision | A process map and a rule for when to stop or escalate |
These are examples of demonstrable skills, not a ranking of what every employer wants. If a role asks for machine learning engineering, ordinary chatbot use is not a substitute for the technical requirements.
Where should AI skills go on a resume?
Experience: use this when the work happened in a real job or volunteer role. Explain what you contributed to an actual task.
Projects: use this for a course exercise, student activity or independent demonstration. Include the project name, what it does and a link if you can safely share one.
Skills: use a short supporting list of methods and relevant tools you can demonstrate. The experience or project section should give that list substance.
Education or training: use this for completed courses. Keep a completion certificate separate from a professional certification or an accredited qualification.
The University of Pittsburgh's career guidance also recommends explaining the problem you solved, your decisions and your contribution, rather than listing tools without context.
A simple formula for an honest CV bullet
Use this structure:
[Action you took] + [specific task] + [how AI helped] + [your check or documented result].A number is optional. If you did not measure time saved, errors reduced or sales increased, do not invent a percentage. A precise description of the deliverable is better than a made-up improvement.
Example 1: an operations coordinator preparing a team briefing
Too vague: "Skilled in AI and ChatGPT."
Illustrative alternative: "Used an approved AI assistant to turn permitted project notes into a team briefing; checked each action owner, deadline and open decision against the source before sharing the final version."
This makes the source material and the human check visible. It describes drafting and review without implying that the coordinator automated project management or developed a model.
Example 2: an independent customer-support project
Too broad: "Built an automated customer service system."
Illustrative alternative: "Created a practice workflow for drafting replies to fictional customer enquiries, with a human review step for refunds and missing order details."
The word "practice" matters. If the workflow never handled real customers, it should not be described as a deployed business system.
Example 3: a professional comparing software options
Unproven: "Used AI to improve research efficiency by 80%."
Illustrative alternative: "Prepared a comparison brief from public product documentation using AI-assisted summarisation; checked feature claims against the original pages and recorded access dates."
This describes the work without pretending a time-saving experiment happened.
These examples are invented to show the writing pattern. They are not learner testimonials or results from a study. Adapt a sentence only if it accurately describes work you have done.
What if you cannot share work from your job?
An employer's document does not become suitable for a public portfolio just because you remove its logo. Keep restricted material in approved systems. You can describe your method at a permitted level or create a separate demonstration with fictional inputs. Label a demonstration as practice, even when it resembles a task you do at work.
Here is a small exercise for an internal workshop FAQ:
- Write a fictional team workshop brief: a date, location, attendance rule and contact method.
- Ask an AI assistant to draft five attendee questions and answers using only that brief.
- Add one question the brief cannot answer, such as whether travel costs are reimbursed.
- Check whether the tool invents a policy or clearly says the information is missing.
- Revise the prompt and edit the final answers yourself.
- Save the brief, both drafts, your corrections and a short account of the limitation you found.
That collection shows more than a screenshot of an impressive-looking answer. It lets a reader inspect your decisions.
Use the AI project evidence template to organise the files. For a longer guided version, take the free AI portfolio course.
Ask AI to check your wording, not invent your experience
Use a tool and inputs permitted by your employer. Removing names alone may not make internal material safe to share. For a public chatbot, use an independent example with fictional or public information, and leave personal contact details out.
Help me describe a project honestly for a job application.
Target role: [role and relevant requirements]
What I actually did: [your notes]
How AI helped: [the specific step]
What I checked myself: [your checks]
Evidence I can show: [files or a public link]
Measured results, if any: [leave blank if not measured]
Suggest two concise CV bullets using only these facts.
Do not invent numbers, responsibilities, job titles or qualifications.
Flag anything that needs more evidence and ask me about it.Read the result as a claim about yourself. If it makes your contribution sound larger than it was, edit it. You are responsible for the final wording.
Be ready to explain the work without the tool
Before adding a bullet, practise answering these questions:
- What was the original problem, and why did you use AI for it?
- What did the tool produce, and what did you change?
- Which error or limitation did you find?
- How did you verify the final result?
- What would still need a person's decision?
If you cannot answer one of these, return to the project. The missing explanation is useful feedback about what to practise next.
Your final check before applying
Keep the bullet relevant to the job, label practice work clearly, remove unsupported numbers and confirm that you can show the evidence without sharing private material. Follow any application instructions about AI use. No wording trick can guarantee an interview or replace the employer's requirements.
For help editing the whole document, read Can AI Write My CV?. To build skills through everyday tasks, use the professional learning path. If you are preparing for a move, our guide to AI skills that help you demonstrate useful work connects those skills to job applications.