Using AI for your resume, honestly

What AI is genuinely good at

• Rewording a clumsy sentence so it reads clearly.
• Choosing what to emphasize for one particular job.
• Structuring a resume and keeping its headings and dates consistent.
• Pointing out what a posting asks for that you have not mentioned — as a prompt to remember something true, not as a licence to add it.
• Drafting a first cover letter and practising interview questions.

Where it goes wrong

An AI tool writes what sounds right, and a resume is full of places where a plausible sentence is not a true one. These are the six patterns we found most often while testing our own tool:

1. Duties by association. You list “networking” as a skill; the draft says you “configured firewalls”.
2. An invented “how”. You cut downtime from 95 minutes to 40; the draft adds “through improved runbooks”. The number is yours. The method is not.
3. Wider scope. “Supported three sites” becomes “across the entire organization”. Adjectives such as “high-volume” or “enterprise-scale” slip in the same way.
4. Upgraded ownership. “Helped with the budget” becomes “led the budget process”.
5. The posting, echoed back. The job asks for experience with a tool you have never used, and the draft quietly writes it as something you have done.
6. A cover letter that repeats the embellishment. Written from the resume rather than from your record, the letter treats the resume’s inventions as established facts.

Careful tools reduce these; none removes them. Our worked example lists the lines our own checking missed, in the open.

A method that keeps you honest

1. Write the facts first. Roles, dates, tools, results with numbers. This is your record.
2. Give the AI the facts and the posting, and tell it to use only the facts. A prompt that works:

   Rewrite these bullets for the attached job posting. Use only the facts I have given you. If the posting asks for something I have not mentioned, do not add it: list it separately as a question for me.

3. Treat every addition as a question, not an upgrade.
4. Read each line against your record. Highlight any claim you cannot point to evidence for.
5. For each highlight: delete it, fix it, or turn it into interest. “Keen to learn Procore” is honest. “Experienced with Procore” is not.
6. Say it aloud. If you could not explain a line in an interview with an example, it should not be on the page.

Words that deserve a second look

• Ownership verbs: led, managed, owned, spearheaded.
• Sweeping words: all, every, entire, across the organization, high-volume.
• Any number or percentage you cannot trace to a record.
• Level words: expert, advanced, extensive.
• Tool names copied from the posting.
• “Current” and “certified”: certificates expire.

Privacy: before you paste

A resume is full of personal information. Before using any AI tool, find out four things: who receives what you type, whether it is used to train models, how long it is kept, and how to delete it. If the answers are not published, ask. As an example of what a clear answer looks like, here is how we answer them.

Also: do not paste a current or former employer’s confidential information into any AI tool, and leave out what a resume never needs, such as your social insurance number or date of birth.

Employers and AI

Some employers ask applicants about their use of AI, and some have rules about it. Follow what the application says. Either way, what matters in the end is that you can stand behind every line when someone asks you about it.

Five questions before you send

1. Can I explain every line in an interview?
2. Is every tool, certificate and duty on the page something I really have?
3. Do the numbers match my records?
4. Does the cover letter say anything the resume does not support?
5. Have I read every word myself, not just skimmed it?

Print the checklist to keep beside you.

Published by My Resume Robot. Last updated October 15, 2026.