ResumeAI — State of ATS 2026: No market need

ResumeAI — State of ATS 2026 was a free ATS-safe resume builder bundled with a hand-verified dataset mapping 743 large employers to the applicant tracking systems they actually run. The founder is listing it not because the site is dead — it runs live with automated weekly re-verification and two paying customers this month — but because a new full-time job leaves no capacity to grow it, and the consumer resume tool never found product-market fit.

The Resume Builder Was a Commodity Trap

The consumer product offered nine templates, AI bullet generation, cover letters, interview prep, a job tracker, a resume checker, and a Chrome extension. Building was free; a clean, watermark-free download cost $12.99 one-time or $19.99 per month. Over the last 30 days the funnel saw roughly 34 real humans per day, 179 resumes built, 49 signups, four checkout starts, and one completed payment. Lifetime revenue across three customers totals $14.97.

The structural problem is that job searches end. There is no retention loop to compound: a user builds a resume, gets hired, and leaves. Incumbents win on a decade of SEO authority and subscription models that rely on users forgetting to cancel. The founder tested one-time pricing, a $0.99 first month, and 24-hour reverse trials — 63 users entered the reverse trial and zero converted. A stronger download watermark, personalized paywalls showing the target company's verified ATS, exit-intent capture, and lifecycle recovery emails were all tried. Only the lifecycle email produced a sale, recovering a Bing visitor who subscribed at $19.99 per month.

The Dataset Was the Real Product — And It Proved Its Own Value

State of ATS 2026 maps 743 Fortune 500 and Global 2000 employers to their live ATS. Of those, 707 were verified by hand against the apply-URL host on the employer's careers portal in June 2026 and fully re-verified in July. An extended tier adds 279 machine-verified employers. The data is MIT-licensed, free on purpose, mirrored to GitHub, npm, Kaggle, and Hugging Face, and served through a public JSON API, an OpenAPI spec, and a remote MCP server for AI agents.

Automated monitoring re-probes roughly 550 of those employers weekly against live vendor infrastructure. Detected changes go to human review before publication. Confirmed migrations publish to a public changelog and RSS feed with a curl-able evidence host on every entry. The July re-verification found that approximately 24 percent of the hardest-to-check records had rotted in a single month: 14 acquired brands still listed as independent, two defunct employers, and roughly 60 silent vendor migrations. Taleo alone dropped from 39 employers to 18.

That rot rate is the commercial pitch: every static ATS dataset is wrong within a quarter, and this one re-checks itself. A Harvard-affiliated health writer cited the report unprompted. A Wall Street Journal reporter asked to see it. An auto-apply SaaS CTO took a call. The dataset — built originally as a side project to make the resume builder smarter — became the only asset anyone wrote about, replied to, or asked for.

Distribution Failed Until AI Assistants Showed Up

The founder spent months confusing shipping with distribution. Google never indexed the site meaningfully; the domain sits at DA-15. Roughly 120 cold B2B emails produced zero sales. A $299 self-serve snapshot and a $199 per month feed also produced zero.

The only two revenue events came from surfaces the founder did not control. A Microsoft Copilot recommendation converted in six minutes. A Bing visitor recovered by a lifecycle email became the subscription customer. Copilot referrals grew from four to six to 16 per month over recent months. The AI-referral channel is small but real and growing.

The infrastructure to capture that channel cost a weekend: an llms.txt file, structured data, an OpenAPI spec, and the remote MCP server. The first full-price sale arrived through Copilot within six minutes of landing. Being the source AI assistants cite is an underrated distribution channel in 2026, and almost no data publishers have wired themselves for it.

Cold Outbound Into an Unbudgeted Category Is a Waste of Time

The B2B target — auto-apply agents, job-data vendors, unified-API platforms, HR-tech marketers — had not budgeted for an employer-to-ATS attribution feed. One hundred twenty cold emails, a priced snapshot, and a priced feed yielded nothing. One inbound citation from a journalist did more for credibility than the entire outbound campaign.

The lesson is not that cold email never works. It is that cold email into a category the buyer has not yet lined up budget for rounds to zero. The dataset solves a real problem — silent ATS migrations break auto-apply workflows and corrupt job-data pipelines — but the buyers who feel that pain acutely are not yet shopping for a vendor. They are patching it internally.

Pricing Experiments Couldn't Fix an Intent Problem

The founder ran a disciplined set of pricing tests: $12.99 one-time pass (the current primary offer, which produced the two recent sales), $0.99 first month, 24-hour reverse trials, personalized paywalls, exit-intent capture, and lifecycle emails. The honest read is that the gate is not the problem. Arriving with intent is.

Visitors who came from an AI assistant recommendation converted. Visitors who came from search-engine listicles did not. The funnel — 179 builds to 49 signups to four checkout starts to one payment — shows that users extract value from the free on-screen preview and leave. They do not churn; they never convert. Job search is a finite episode. Subscription billing fights the user. One-time pricing on a commodity tool is still a commodity tool. The mistake was assuming pricing was the lever when the real problem was that almost nobody arrived wanting the product.

What a Buyer Gets

The sale includes the full Next.js (App Router) codebase with TypeScript, Tailwind, and the Chrome extension (MV3), deployed on Vercel with Supabase (Postgres, Auth, Storage), Stripe billing, Anthropic Claude API integration, PostHog analytics, and Resend email. The live dataset — 743 employers mapped to ATS, 707 hand-verified, 279 machine-verified — plus the automated weekly re-verification pipeline (roughly 550