How to list ChatGPT Work on your resume in 2026

What does Uber putting 30,000 employees on ChatGPT Work mean for your CV?

On 22 October 2026, Uber said it had moved its entire global workforce — about 30,000 people — onto ChatGPT Work, scaling up from a pilot of roughly 1,000 employees. For your CV, that matters in one very specific way: thousands of people acquired the same line of experience on the same day, so the line itself is worth nothing. What you did with the tool is worth plenty. Uber's engineering leadership said around 80% of early pilot users were non-engineers, spread across finance, operations and product, and that the company would track adoption and workflow outcomes rather than publish a headline productivity figure. Read that as your template too: employers care about the workflow you changed, not the seat you were assigned.

ChatGPT Work launched in July 2026 as an agent inside ChatGPT Enterprise, built to take a multi-step goal and hand back a finished artefact — a report, a deck, a spreadsheet, a cleaned-up data pull — using context from connected tools like Slack, Google Drive, Gmail and the calendar. Uber's rollout kept software engineers on its Codex-based setup, paired roughly 30 AI engineers with business teams for two-week sprints, and shipped an internal library of AI skills alongside training. Every one of those details is CV material for somebody: the person who ran a pod, the person who wrote a reusable skill, the person who trained a department. The subscription isn't the story. What got built on top of it is.

Here's the part most job seekers get wrong. Demand for AI capability in job ads has climbed hard — US postings asking for AI skills were up well over 100% year on year by spring 2026, and more than a third of entry-level roles now expect some AI competence — yet the brand names barely appear. Large scans of US listings put ChatGPT itself in only a single-digit share of ads, while generative AI, AI integration and prompt design show up far more often. Employers are buying capability, not your login. Which is why "proficient in ChatGPT Work" matches almost nothing in an applicant tracking system and tells a recruiter nothing they hadn't already assumed about you.

Where should ChatGPT Work go — the tools line or an achievement bullet?

ChatGPT Work belongs in exactly two places on your CV: once in a tools line inside your skills block, and once inside an achievement bullet in the role where you actually used it. The skills line exists for matching, because recruiters and screening filters search literal strings and the words need to sit in plain, extractable text. The bullet exists for credibility, because a tool name on its own proves nothing at all. Anything beyond those two mentions starts to read as padding. If you genuinely owned a deployment rather than used one, you've earned a third mention in your professional summary — but only if you can back it immediately with scope and a number.

Put it where a parser can actually read it. Applicant tracking systems pull raw text out of your file, so a tool name sitting inside an icon row, a sidebar or a header text box can come back empty — and a recruiter running a keyword search will never see you. A plain line does the job: "AI and tools: ChatGPT Work, Slack, Google Workspace, Looker, SQL." Notice that it sits next to the rest of your stack instead of in a separate AI cloud with logos. If your current template leans on two columns or decorative blocks, rebuild it in a clean single-column format before you spend another minute on wording.

Three habits sink this section. First, the standalone claim "AI literacy" — it's the new "familiar with Microsoft Office", and experienced recruiters skim straight past it. Second, the twenty-tool list: it signals dabbling, and it looks copied from a listicle rather than earned at work. Pick the two or three tools you touch weekly and can defend. Third, the summary buzzword, as in "AI-driven professional passionate about transformation" — that's a sentence with no information in it. A useful filter before anything goes on the page: could you answer what problem you solved, what you personally did, and what changed as a result? If not, cut it.

How do you turn "we rolled out ChatGPT Work" into a measurable bullet?

Rewrite it with three components: the workflow you changed, what you specifically did, and one number that moved. "We rolled out ChatGPT Work" is company news. "Rebuilt the weekly ops reporting pack in ChatGPT Work, cutting prep from six hours to ninety minutes and freeing two analysts for forecasting" is your work. The verb matters as much as the number — built, rewrote, trained, automated, documented and standardised all describe something a person did, while used, helped with and was involved in describe something that happened nearby. If a hiring manager can't tell from the bullet whether you drove it or watched it, they'll assume the second one.

Four numbers are usually available to you, even if nobody formally measured anything. Cycle time: how long a recurring task took before and after. Volume: tickets, reports, briefs or candidates handled per week at the same headcount. People enabled: how many colleagues you trained, or how many teams picked up a workflow you wrote. Quality: error rate, rework, QA pass rate, first-response time. Reconstruct the "before" honestly from your calendar, your ticket queue or a monthly report rather than guessing. An approximate, clearly bounded figure — "roughly 15 hours of manual data entry a month" — is far stronger than a precise-sounding invention you can't defend when someone asks.

Attribution is where these bullets live or die. Say what the agent produced and what you owned: the brief, the data checks, the review gate, the final call. "Drafted with an agent, verified against the source system, approved by me before publication" is the shape hiring managers trust in 2026, because they've all seen confident AI output that was quietly wrong. Managers now say AI makes individual contribution harder to judge, so the candidate who separates the two reads as senior. There's more detail on this in our guide to writing impact bullets when AI did part of the work, including which claims to keep for yourself.

What's the difference between using an AI tool and owning a deployment?

Using a tool and owning a deployment are different jobs, and blurring them is the fastest way to lose an interview you'd otherwise have won. There are roughly three honest levels. You're a user if you work inside the agent daily and have made your own tasks faster. You're a workflow builder if you've designed something repeatable that other people now run — a documented prompt set, a connected pipeline, a template your team relies on. You're an owner if you were accountable for rollout: scope, enablement, data rules, approval gates and measurement. Each level deserves different language, and the top one deserves your summary line.

An owner's bullet reads differently because it carries the unglamorous parts. Headcount and footprint: how many people, how many teams, which countries. Enablement: training sessions delivered, champions recruited, a skills library written. Governance: what data was allowed anywhere near the tool, which outputs needed human sign-off, what was explicitly off-limits. Measurement: what you tracked, and what you decided when adoption stalled in one department. That last item is the one interviewers remember, because every real rollout stalls somewhere. Uber's own approach — embedding AI engineers with business teams for two-week sprints and shipping an internal skills library — is exactly this kind of work, and it's worth describing in that concrete a form.

Resist the temptation to promote yourself a level. Interviewers probe AI claims hard now, and the question that collapses an inflated bullet is embarrassingly simple: "walk me through what you changed in week one." If you were a keen user, say so with a strong number and you'll still beat the candidate claiming to be a transformation lead with nothing underneath it. One contrarian note: for most non-technical roles, a single deep, verifiable workflow beats breadth every time. If you're writing up agent-based work specifically, our breakdown of how to phrase agent oversight on a CV covers the titles and bullet formulas that hold up under questioning.

Which AI phrases get cut, and how do you check your CV survives screening?

Cut anything that could appear on any CV in your industry without changing meaning: "AI-powered mindset", "harnessed the power of AI", "prompt engineering expert", "passionate about emerging technology", "AI enthusiast". Recruiters read those as filler at best and as AI-generated text at worst. Keep the specifics instead — the tool, the task, the artefact, the number. "Wrote the agent workflow our three regional support teams now use for tier-1 replies" survives scrutiny; "driving AI transformation" does not. The test is whether a sceptical stranger could picture what you did on a Tuesday afternoon. If they can't picture it, a hiring manager can't either.

Then tailor it to the ad in front of you. Mirror the employer's own vocabulary: if the posting says "generative AI", don't write "GenAI" and hope the filter is clever. If it says "AI agents" or "automation", use those words in a bullet where they're true. Ads increasingly ask for AI competence in marketing, HR, finance and operations roles, not just engineering, and the phrasing varies wildly between companies — so one generic AI paragraph reused across fifty applications will underperform three tailored versions. Keep a master CV with every quantified AI bullet you own, then cut down to the two or three that match each role.

Finally, check the file rather than trusting your eye. A free analysis at HRLens scores your CV out of 100, breaks it into five categories — experience, tech stack, impact and ownership, clarity and structure, ATS compatibility — and shows a visual layout analysis of how your design reads once the text is extracted. Paste the job description and you'll also see which skills the ad asks for that your CV never mentions. If your impact and ownership score comes back low after you've written up a company-wide AI rollout, the problem isn't your experience. It's that the page is still describing the tool instead of you.

Frequently asked questions