Six in ten job postings mention AI, but only about a third ask the candidate to use it. What the asks look like by role, and how to show AI skills on a resume.
Published by MatchCV. Editorial scope and corrections.
Every resume guide published this year says to add AI skills. Before adding them, it's worth knowing how many postings ask for them, and what "ask" means in the posting's own words.
We counted across the same set of postings as our job posting keyword study: public openings from 89 companies, collected on 14 October 2026. The short version is that most postings mention AI, a minority ask you to use it, and the two numbers are far enough apart to change your advice.
Our first pass counted any posting containing "AI," "LLM," "machine learning" or similar. It came out at 85% of postings, including recruiters and sales reps. That number was wrong, and it's worth explaining why, because most "X% of jobs want AI" statistics have the same flaw.
Company boilerplate was the cause. Every posting from one data company opens with "the Data and AI company." Every posting from one restaurant-software company has a paragraph about how it uses AI. Once we removed text repeated across a company's roles, the figure dropped to between 42% and 72% depending on the role. Most of what remained was still product description, such as "AI-powered platform" or a slogan.
So we measured something narrower: a sentence that tells the candidate to use, leverage, be fluent with, or have experience with AI. We skip a match when the words before it are about customers or the company, and when the AI term is just a product modifier ("AI-powered platform"). Compare the two kinds of sentence.
We also left out the boards of companies whose product is AI itself, such as Anthropic, OpenAI and Perplexity, because nearly every posting there asks for AI by definition.
How reliable is the narrow measure? We read 60 randomly chosen matches. 55 were real asks to the candidate. Of the five errors, three were company descriptions such as "AI-first environment," which came from one loose rule that we then removed, and two were product duties ("using AI to reduce friction, predict drop-off"). That's roughly nine in ten, so read the percentages below as approximate and expect them to run a little high.
| Role | Postings counted | Mention AI anywhere | Ask the candidate to use or know AI |
|---|---|---|---|
| Security engineer | 159 | 63% | 32% |
| Product manager | 408 | 72% | 30% |
| Software engineer | 943 | 63% | 30% |
| Program manager | 186 | 57% | 28% |
| Sales rep | 578 | 42% | 13% |
Postings counted are those left after excluding AI-first company boards. Percentages use at most 20 postings per company and role.
About three in ten postings in the technical and product roles ask for it. For sales it's about one in eight. Around seven in ten of the sales postings that mention AI don't ask the candidate to use it at all.
We sorted the asks into three kinds, using the words around the AI term. The split is approximate, since many sentences don't fit one box.
| Role | Using AI tools in daily work | Building with AI (LLMs, agents) | Both | General ("familiar with AI") |
|---|---|---|---|---|
| Software engineer | 56% | 15% | 9% | 20% |
| Product manager | 50% | 21% | 4% | 25% |
| Security engineer | 26% | 41% | 2% | 31% |
Shares are of the postings that ask. For software engineers, the typical ask is about how you work: "Use AI-assisted developer tools to move faster" or "Fluency with AI-assisted engineering." Far fewer ask you to build LLM features, and named machine-learning libraries are rare in ordinary engineering postings. PyTorch and TensorFlow appear in under 3% of software engineer postings, although "LLMs" appears in 19% and "machine learning" in 17%.
Security engineer postings lean the other way. They're likelier to ask about AI systems themselves, such as securing models or AI-related threats, and "AI security" turns up in 13% of them.
The data suggests three things.
Describe a habit, not a keyword. Postings phrase the ask as a way of working: use AI in your daily work, with judgment. A skills line reading "AI, LLMs, prompt engineering" answers a question the posting didn't ask. A bullet does better:
Only write the second version if it's true of you. The invented numbers in that example are placeholders for facts you'd have to supply from your own work. Interviewers do ask follow-ups on this, and postings that say "with human judgment in the loop" are telling you what the follow-ups will be.
Match the kind of ask. If the posting is about daily tooling, name the tool and the task. If it's about building, name the system you built. If you run a security team, the credible line is about securing or evaluating AI systems, not about autocompleting code.
Don't add it where the posting doesn't ask. For a sales rep or recruiter posting that never says "use AI," an AI skills line takes space from things the posting does say. The keyword scanner shows which terms a posting contains, so you can see whether yours is one that asks.
If you're deciding how much of your resume to rewrite around AI, start from your own target postings. Paste one next to your resume in the keyword scanner and see whether it asks at all. The step-by-step version is in our guide to tailoring a resume in five minutes.
Find the keywords your resume is missing for any job description
Scan my resume keywords →or sign in free to tailor your resume with AI