Do AI skills on a resume help outside tech?

What did the 'no AI payoff' report actually say?

On 24 October 2026, Apollo Global Management's chief economist, Torsten Slok, published a short market note under a blunt headline: no signs of AI payoff outside the tech sector. His argument is that despite enormous investment and relentless hype, AI has delivered zero financial payoff to corporate America beyond technology companies. The evidence is narrow but hard to wave away. Strip the seven giants — Alphabet, Amazon, Apple, Meta, Microsoft, Nvidia and Tesla — out of the S&P 500, and the remaining 493 companies show profit margins parked near 10% and revenue per worker that is essentially flat, while the big seven's revenue per worker keeps climbing. Executives name-drop AI on every earnings call. The margins haven't followed.

Keep it in proportion: this is a short macro note built on charts, not a controlled study, and it can't prove AI fails inside any individual company. What it does show is where the money goes. Slok's figures put the top 10% of firms behind 99.5% of spending on model serving and inference, and more than 95% of the spend tied to the big model providers. Everyone else, he argues, has begun treating token spending as a cost center that drags on margins rather than a growth engine. That's a sentiment shift, and sentiment is exactly what sits across the table from you in an interview. A finance director who already wrote off one AI pilot reads your claims differently.

Here's the bridge to your CV. If you're applying to a bank, a hospital group, a school district, a logistics operator or a mid-size manufacturer, the person reading your application has probably sat through an AI pilot that produced a slick demo and no savings. A skills line reading "ChatGPT, Claude, Gemini, Copilot" tells that reader you spent time and license budget. It says nothing about return. The candidate who writes "cut invoice-matching time from 14 minutes to 4 by adding an AI review step, clearing month-end three days early" is answering the precise doubt the note describes. Same technology, opposite signal. That gap between tools listed and results proven is the whole subject of this article.

Do AI tool names on a resume still earn interviews?

No — not on their own, and not since about mid-2025. Naming ChatGPT, Copilot or Claude in a skills row has become table stakes rather than a differentiator. Resume samples tracking AI-related terms found that the share of resumes mentioning at least one AI term climbed from 3.7% in 2023 to 12.8% in 2025, roughly one in eight. Any signal that jumps that fast stops distinguishing anyone. Recruiters have learned to read a bare tool list the way they read "Microsoft Office": proof you've been alive recently, not proof you can do the job. The claim now has to carry a consequence attached to it.

Look one level down and the picture gets more useful. In the same sampling, the generic terms are the ones that exploded — "artificial intelligence" went from 0.5% of resumes to 6.3%, "machine learning" from 0.6% to 5.7%. The specific, operational language stayed rare: generative AI terms reached only about 1.5%, ChatGPT and OpenAI around 1%, and prompt engineering just 0.6%. Most individual AI terms still appear on under 5% of resumes. So the crowding is at the buzzword end. If you can describe a concrete workflow you built and the number it moved, you're competing in a much emptier field than the headline percentages suggest.

A contrarian suggestion: delete your "AI Tools" skills row entirely and spend those characters inside your experience bullets. Keyword matching in an applicant tracking system is a crude gate — the system scores whether terms from the job ad appear in your document, and it doesn't care whether the term sits in a skills list or a bullet. The human who reads you next does care, enormously. A tool name inside a results bullet clears both hurdles; the same name sitting alone in a comma-separated list clears only the first and then quietly costs you credibility with the second reader.

How do you turn AI work into a measured business result?

Pick a number the business already tracks, capture it before and after, state the scope, and mention the tool only as the mechanism. That's the entire method. Finance teams track days to close and error rates. Support teams track first-response time, handle time and reopen rates. Marketing tracks cost per lead and production cycle time. HR tracks time to fill and screening hours. Operations tracks rework, scrap and on-time delivery. You don't need a company-wide figure — your own desk is a legitimate unit of measurement. "Reduced my weekly reporting block from six hours to ninety minutes" is small, specific and instantly believable, which is worth more than a vague claim about transforming a department.

Write it in four parts: the outcome, the number, the scope, and the mechanism in a short trailing phrase. Outcome first, because that's what gets read. Scope matters more than people think — "across 40 weekly invoices" or "for a 12-person team" turns a percentage into something a hiring manager can picture, and a percentage without a denominator reads as padding. Be honest about attribution: if you were one of five people on a project, write "contributed to" and then claim the slice you personally owned. Overclaiming is the fastest way to lose an interview, because the follow-up question is always "walk me through how you measured that."

The rewrite below shows the pattern across four non-technical functions. Notice that the strong version never opens with the tool, and in two cases the tool isn't even named — the process change is the story. If you're rebuilding bullets from scratch rather than editing, a chat-based CV builder that drafts and refines your bullets will get you to a tidy, parseable draft faster than fighting a template in Word, and you can keep reshaping each line until the numbers sit at the front. Either way, write the number before you write the sentence around it.

Which AI skills actually count for non-tech roles?

The ones tied to a controlled process: verifying AI output, automating a workflow end to end, reading data well enough to challenge a model's answer, and knowing the governance rules that apply to your industry. Employers outside engineering aren't shopping for people who can fine-tune models. They want someone who can decide where AI belongs in a process, measure whether it helped, and catch it when it's wrong. That last one is the quiet differentiator. "Reviewed and corrected AI-drafted client summaries, with a 2% error rate at monthly audit" describes a skill a compliance officer actually values, and almost nobody writes it that way.

The demand is real but smaller than the hype suggests, which is why specificity pays. Stricter measures of job postings put explicit AI skill requirements in roughly 2.5% of US listings in 2025, up about 55% year over year, while roles asking for at least one AI skill grew around 69% during 2025 against roughly 9% for jobs overall. Crucially, about half of AI-skill postings sit outside IT. The fastest-rising non-engineering capabilities skew strategic and procedural: AI business strategy, data governance, responsible AI use, process optimisation and workflow automation. Those are the phrases a non-tech hiring manager recognises from their own board deck.

Placement, then. Put one positioning line in your summary — something like "operations lead who rebuilt two reporting workflows around AI review, cutting reporting hours 60%" — and keep a short grouped line in skills, "AI and automation: workflow design, output verification, data governance," so the parser finds the terms. Then let your bullets carry the proof. Don't claim a capability you can't demonstrate in ten minutes: writing "machine learning" when you've only used chat tools invites a technical question you'll fail. And skip the certificate wall. One relevant short course is fine; four of them reads like compensation for missing experience.

Does your resume read as proof or as hype?

Run one test on every AI claim in your document: could a stranger verify it or argue with it in an interview? Proof has a number, a timeframe, a scope and a named system. Hype has an adjective. "Drove AI-powered transformation across the organisation" survives no follow-up question at all, while "reduced manual data entry by 22 hours a month across a 5-person billing team" invites exactly the conversation you want. Read your bullets aloud and mark every sentence where the only concrete noun is a tool name. Those are the lines costing you interviews, and in a market that's just been told AI hasn't paid off, they actively work against you.

The reliable hype markers are easy to spot once you look for them: "cutting-edge", "AI-driven", "next-generation", percentages with no baseline, improvements with no denominator, and verbs that describe enthusiasm rather than action. Inflated attribution belongs on the list too — claiming a department-wide result you touched for two weeks. There's a related worry worth separating out: sounding like a machine wrote your CV is a different problem from overclaiming AI skills, and the fixes differ. If you're unsure which one you have, it's worth reading up on the phrasing that makes a draft look AI-generated before you send anything.

You can also get the judgement done for you in a couple of minutes. A free AI CV analysis returns a score out of 100 plus five category scores, and the one to watch here is impact and ownership — it reads whether your bullets carry measured outcomes you clearly owned, or whether they describe activity. You'll get ATS compatibility scored alongside it, and a visual layout check. Paste the job ad text in too, and the analysis lists the skills you're missing against that specific role, then generates ATS-friendly rewritten CVs in six templates that mirror the ad's language. Separately, if you're wondering whether to mention using AI at all, the rules on disclosing AI help are narrower than most people assume.

Frequently asked questions