Accounting
Weak
Used ChatGPT to improve monthly reporting.
Evidence-based
Redesigned a monthly close checklist with AI-assisted gap detection; retained controller approval for controls and tracked reopened ownership questions across three closes.
AI skills for Canadian professionals
Show the workflow, your judgment, and the measured result.
A résumé line that says “ChatGPT” names access to a tool. A stronger line explains the work you changed, the review you kept, and what improved after you tested the workflow more than once.
Four parts of a defensible claim
Write down the evidence before compressing it into one bullet.
- 01
Name the work
A repeated task such as meeting actions, acceptance criteria, close steps, or a client brief.
- 02
Bound the AI role
Drafted, grouped, compared, or surfaced questions; it did not own the final decision.
- 03
Name your review
Verified sources, facts, calculations, security, fairness, commitments, or professional requirements.
- 04
Show the result
Time to a reviewed output, correction rate, reopened work, missed dependencies, or another recorded outcome.
Before and after
Five AI résumé examples that show work rather than tool access.
Consulting
Weak
Experienced with AI research and proposal writing.
Evidence-based
Built a discovery-to-brief workflow that separated client statements from hypotheses and required source and commitment review; tracked drafting time and client corrections across three briefs.
Management
Weak
Used Copilot to increase team productivity.
Evidence-based
Introduced a reviewed meeting-to-action workflow; verified every owner and deadline and measured preparation time and follow-up corrections over four weekly meetings.
Human resources
Weak
Used AI in recruiting and onboarding.
Evidence-based
Built a 30-day onboarding sequence from approved role requirements; retained manager review and tracked missed dependencies without sending employee records to the AI tool.
Software engineering
Weak
Proficient with ChatGPT and AI coding tools.
Evidence-based
Designed an AI-assisted acceptance-criteria workflow that surfaced ambiguity before build; retained security, accessibility, feasibility, and product-intent review and tracked reopened work.
Evidence does not require exposing employer information.
Describe the method and result at the level your employer or client permits. Rebuild the demonstration with public or synthetic material when the original input, prompt, output, or metric is restricted.
Check the workplace data boundaryA sanitized work sample should still show
- The original task and the reason it mattered
- A public or synthetic example of the input and output
- The checklist you used to review AI-assisted work
- A measured result or an honest statement that the test is not complete
The interview story needs one more layer than the résumé bullet.
Situation
What repeated work was slow, inconsistent, or difficult to review?
Boundary
What information and decision did you keep outside the AI tool?
Review
Which errors, assumptions, or risks were you personally responsible for catching?
Result
What changed across repeated tests, and what would make you stop using the workflow?
Questions people search
Listing AI skills on a Canadian résumé
Should I list ChatGPT or Copilot as a skill on my resume?+
List a tool only when it matters to the target role and you can explain how you used it. The stronger evidence is a work task, your review responsibility, and a result that can be checked.
How do I describe generative AI experience on a Canadian resume?+
Write the bullet around the business workflow: what you changed, what AI assisted with, which decisions remained human, and what improved across repeated use. Keep employer and client information confidential.
Can I put a personal AI project on my resume?+
Yes, when it produces a work sample a reviewer can check and uses public, synthetic, or authorized information. Label the context honestly and do not present a personal exercise as employer deployment experience.
What if I cannot share the AI work sample?+
Describe the workflow, review method, and result without disclosing restricted details. You can also rebuild the example with synthetic material and state that it is a sanitized demonstration of the method.