Should you list a dead AI tool on your resume?

Should you keep a shut-down AI tool on your CV?

Keep the skill, drop the dead brand name — with one exception: if the shut-down tool is still named in the job ad, keep it and date it. The question got sharper on October 24, 2026, when the running tally of AI projects that didn't make it was updated again. The list is long. Relay, an AI workflow automation startup pitched as a Zapier alternative, closed after five years. Huxe, an audio app from ex-NotebookLM engineers, shut down in May 2026. The model-comparison playground Yupp closed in March 2026 for want of product-market fit. Notion Mail was switched off on September 22. Hardware took hits too, from the Rabbit R1 to the Humane AI Pin.

Those names are sitting on CVs right now — in skills lists, in bullet points written while the product still had a roadmap, in portfolio lines that made someone look early and sharp eighteen months ago. And this isn't a handful of unlucky people. The share of companies that abandon most of their AI initiatives climbed from 17% to 42% in a single year, with the average organization scrapping roughly 46% of its proofs of concept before anything reaches production. Add the model retirements that run on a published schedule — DALL·E 2 and 3 pulled from the API in May 2026, Firebase Studio winding down toward March 2027 — and a big slice of recent AI experience is now attached to something that no longer exists.

So the real risk isn't embarrassment. It's that the line has quietly become a dead keyword: a term nobody puts in a job ad, nobody types into a search box, and nobody recognizes without a pause. A recruiter scanning your CV for eight seconds doesn't think "ah, a casualty of the AI shakeout." They think the stack looks like 2024. That's a different problem from the one most people worry about, and it needs a different fix. If the company itself went under rather than just the product, the employer line has its own rules — there's a separate method for writing a bankrupt employer into your CV without losing the credit for your work.

That's a different problem from the one most people worry about, and it needs a different fix — one aimed at the recruiter's eye, not the parser.

How do ATS keyword matching and recruiters read obsolete tool names?

An ATS has no idea the tool is dead. Applicant tracking systems match the text in your CV against the text in the job ad and the recruiter's search query, so a discontinued product name is neither a penalty nor a pass — it's just a word nobody is looking for. No parser maintains a graveyard list. What you lose is the match you could have had if that same line of space carried a term still in circulation. Think of your skills section as a fixed budget of searchable words. Every slot spent on a product that shut down in March is a slot not spent on the capability a hiring manager is actually filtering for.

The matching itself is more literal than most candidates assume. Named tools, technologies and certifications are typically matched exactly or with light stemming, which is why "Salesforce CRM" and "CRM" don't always behave the same way, and why you spell an acronym out once alongside its short form. Workday, Greenhouse and Lever now run semantic layers that can connect "led a team" to "team leadership" or "Python developer" to "Python engineer." None of them will connect a defunct product to its living replacement. Writing "Operator" will not register as agentic AI experience, and "Figgs AI" will not register as prompt engineering. The machine isn't being unfair — it simply has no reason to know.

Here's my contrarian take: on this particular issue, the ATS is the lesser threat. The bigger one is the recruiter running a Boolean search across their own database, or reading your CV against a requirements list, and finding a stack that reads like a museum exhibit. Dead tool names also invite the wrong question — "wait, didn't that close?" — which spends your credibility on trivia instead of on results. If your employer was absorbed rather than closed, that's a separate fix, and there's precise wording for an acqui-hire on your CV that keeps both company names matchable.

How do you rewrite a bullet tied to a discontinued AI product?

Rewrite it capability-first, number-second, tool-last. The formula is: live skill name → what you built and at what scale → measurable outcome → platform as a closing detail, in parentheses if you keep it at all. "Power user of Relay" becomes "Workflow automation: built 11 multi-step agent workflows for ticket triage and onboarding, cutting manual handling by about six hours a week." The platform may stay as a four-word aside; it is no longer the headline. This single reordering fixes both audiences at once. The parser now reads a term that appears in job ads, and the recruiter reads evidence of judgment rather than loyalty to a vanished brand.

Two rules keep the rewrite honest. First, never drop the employer — a company that filed for bankruptcy is still verifiable employment history, and quietly deleting it creates a gap that costs you far more than the awkward explanation. Character.AI's Chapter 11 filing on October 12, 2026 put a whole team in that position within weeks. Second, keep the numbers that outlive the product. Users served, workflows shipped, latency cut, cost per task, error rate in review, time to launch: none of those stop being true because the tool was switched off. A dead product with live metrics still reads as delivery.

Where the work involved an AI model doing part of the job, be specific about which part was yours — the spec, the evaluation set, the guardrails, the rollback decision. Hiring managers have grown skeptical of bullets that quietly credit the model's output to the candidate, and vague phrasing now reads as a red flag rather than a modern flex. The practical framing for splitting credit between you and the model matters more than the tool name ever did. One more thing: resist the urge to retrofit. If you never touched retrieval pipelines, don't write RAG because it ranks well.

Which AI skills keywords should replace a dead brand name in 2026?

Replace the brand with the capability that outlasted it. The AI terms currently growing fastest in both skill acquisition and hiring success are prompt engineering, retrieval-augmented generation, vector databases, model training and fine-tuning, LangChain, the OpenAI API, data annotation, and evaluation work. Alongside them sits the phrase that has climbed hardest in job ads: practical AI workflow fluency — applying a model to a real task, producing a usable result, and keeping accuracy and compliance intact. Mentions of named assistants in postings have multiplied several times over in a year, while generic "machine learning" references have edged down. Specificity is winning, but it has to be specificity about things that still ship.

Pick the replacement by mapping backwards from what you actually did. Workflow builder on a dead automation platform → workflow automation, agent orchestration, API integration. Character or chatbot work → conversational design, prompt engineering, safety filtering, content moderation. Model-comparison tooling → model evaluation, benchmarking, evals, A/B testing. Image or video generation → text-to-image pipelines, generative media, asset review. Anything involving agents running unattended → agent oversight, human-in-the-loop review, exception handling. That last cluster is the one rising fastest, and it has its own bullet formulas for agent oversight work that are worth copying rather than inventing from scratch.

Then place each term twice, which is how keyword matching actually rewards you. Name the tool or skill once in a dedicated skills section, where parsers read it cleanly, and prove it once inside a role bullet with a number attached. A skills list alone is the most visible and least persuasive place to put a keyword; a keyword inside an achievement carries evidence a recruiter can question you about. If reordering five years of experience around new terminology sounds like a rebuild rather than an edit, it usually is — an AI CV builder that rewrites from your existing file will get you to a clean single-column draft faster than wrestling a template.

How do you check whether your CV still reads as current?

Test it against a live job ad, not against your memory of the market. Take a posting you'd genuinely apply to, list every tool, platform and methodology it names, then check which of those words appear anywhere in your CV — and which of your words appear nowhere in the ad. That second list is your dead-keyword inventory. Anything on it that shut down, got folded into a bigger product, or hasn't shipped since 2024 is a candidate for the rewrite in the section above. Do this with three ads for the same role and the pattern gets obvious fast; one ad can mislead you, three rarely do.

The faster version is to let the analysis run itself. HRLens scores your CV out of 100 for free, with separate category scores for experience, tech stack, impact and ownership, clarity and structure, and ATS compatibility, plus a visual layout analysis that shows whether your formatting even survives extraction. Paste in the job description — copy the text rather than the link for LinkedIn and Indeed, which block fetching — and the analysis lists the skills the ad wants that your CV doesn't mention, then generates ATS-friendly rewritten versions in six templates that mirror the posting's wording. Running your CV through the free analysis takes a couple of minutes.

If you want the harder read, the paid Full Analysis at $3.99 adds the fix list, the "why you might be rejected" breakdown, blind spots, a hiring-probability estimate expressed as interviews per ten applications against the market, and a salary estimate. Either way, treat this as maintenance rather than a one-off repair. AI tools now retire on announced timetables, which means a CV written around product names has a shelf life measured in months. One written around capabilities, scale and outcomes doesn't expire when a startup runs out of runway — and that's the whole argument for how you list AI tools on your resume in the first place.

Frequently asked questions