Every auto-apply tool on the internet claims it gets you more interviews. Almost none of them show their work. This post is us showing ours: the full readout from our beta, including the parts that are shaky.
If you only read one section, read the limitations near the end. The numbers are good. They're also from a small, self-selected cohort over four weeks, and we're not going to pretend otherwise.
The headline numbers
Across a 4-week measurement window with 63 participants:
- Active reApply users shipped 247 applications per month, versus 18 per month for people applying by hand. Roughly 13× the volume, and the user never touches a form.
- The reApply cohort booked 18× more interviews than the control group over the same window.
- Users got back 6 hours a week on average. That's the time the forms, the resume re-uploads, and the "paste your resume, now re-enter it as text fields" ritual used to eat.
The second number is the one we actually care about. Volume is easy to manufacture; any script can submit forms. Interviews resist gaming, because a recruiter has to read what you sent and decide you're worth an hour of their calendar.
How we measured
Who was in the cohort. 63 active job seekers from our beta waitlist, measured over four consecutive weeks. Participants were split between a reApply group, using the product with auto-apply enabled, and a control group that kept running their existing manual search. Everyone was actively hunting. No passive window-shoppers in either group.
What counted as an application. For the reApply group: a submission confirmed by the ATS (Greenhouse, Lever, Ashby, Workday, Workable, SmartRecruiters, or one of the 1,200+ careers pages we support), logged with a receipt. We count confirmations, not attempts. A form that errored out doesn't count. For the control group: self-reported application counts, cross-checked against email confirmations where participants shared them.
What counted as an interview. A scheduled call with a human, whether a recruiter screen or a later stage. Automated one-way video screens didn't count. This is self-reported in both groups, with the same definition applied to each.
What the reApply group was actually doing. Every submission included a resume and cover letter written for that specific role, and only fired when the job's match score cleared the user's floor (users choose between 80% and 95%+). Nobody was spraying. We think that one design choice explains most of the interview number.
Why 247/month isn't "spray and pray"
The obvious objection: of course a bot applies to more jobs. The spray-and-pray crowd has been doing this for years, and their numbers are famously bad. The most-cited Reddit experiment, 5,000 applications over six months, produced 20 interviews. That's a 0.4% callback rate, and the thread under it is full of people reporting the same. We broke down that math in detail in our 30-day auto-apply vs. manual test.
Our beta cohort's behavior looks different in two ways. First, the match floor. reApply only applies when the role actually fits the user's profile above their chosen threshold, so the 247/month average is what qualified volume looks like. These are the jobs a diligent human would have applied to anyway, if they had 40 spare hours a week. Second, every application is bespoke. A tailored resume and cover letter per role survives the ATS keyword screen and still reads like a human wrote it when a recruiter opens it.
The control group's 18 applications a month isn't laziness. It's the realistic ceiling for a person with a job, a life, and a finite tolerance for Easy Apply's broken funnel. The gap between 18 and 247 is just the gap between human hours and machine hours.
The interview multiplier
The 18× figure compares interviews booked by the reApply cohort against the control group across the same four weeks. Three things worth understanding about it.
It's a cohort-level comparison, not a per-person guarantee. Some beta users booked a string of interviews; some booked none in the window. Job markets are lumpy, and four weeks is a short window for senior or niche roles.
Most of the multiplier is volume, and that's fine. If you apply to 13× more well-matched jobs, you should expect interviews to scale up. That's the entire premise. The encouraging part is that interviews scaled faster than application volume (18× vs. 13×), which suggests per-application conversion held up under tailoring instead of degrading the way it does for spray-and-pray.
And the floor matters. A 95%+ match floor means fewer, closer-fit applications; an 80% floor maximizes qualified volume. The cohort wasn't large enough to split conversion rates by floor setting with any statistical honesty. That's one of the questions the next report is sized to answer.
The six hours
Six hours a week is the average time beta users reported getting back. We expected this to be the boring number. It turned out to be the one users talked about most, because the time didn't vanish, it moved. The people booking the most interviews were spending those hours on prep, salary research, and emailing hiring managers directly, a strategy that pairs well with auto-apply. We wrote up the combined playbook here.
The job search has a small core of work that matters (interviews, prep, negotiation) wrapped in a much larger shell of work that doesn't. The reclaimed hours only help if they get reallocated to the core.
What this data can't tell you
We want this report to be citable, which means being upfront about its limits.
- n=63 is small. Big enough to be interesting, far too small to be definitive. The confidence intervals on a cohort this size are wide, and we're not dressing the results up with decimal places they can't support.
- Four weeks is short. It captures application-to-first-interview, not application-to-offer. A longer window could compress or stretch the multiplier.
- Beta users are self-selected. People who sign up for an auto-apply beta are motivated and tech-comfortable by definition. The control group came from the same waitlist to keep the comparison honest, but neither group is a random sample of all job seekers.
- Interview counts are self-reported. We defined "interview" identically for both groups and cross-checked what we could, but self-report is self-report.
- This is one market moment. 2026's hiring market is AI-screened, volume-saturated, and weird. Numbers from this window may not transfer to a different cycle.
What we're tracking next
The next report will follow the cohort further down the funnel: offers, time-to-offer, and how match-floor settings affect interview-to-offer conversion, which is the stage where tailoring should matter most. We'll publish it the same way, methodology and caveats included.
If you want your numbers in the next cohort, the waitlist is at the bottom of this page. We'll do the 240 forms. You do the six calls that matter.