Debugging a Candidate Stuck in the AI Screening Pipeline

A hiring manager pings a recruiter about one specific candidate — someone they met at a placement event, someone they actually want an update on — and the recruiter opens HireQwik and can’t find where that candidate stands. Not rejected, not scheduled, not a verdict. Just… nowhere obvious. “It’s somewhere in the pipeline” is not an answer a hiring manager will accept twice.

This happens rarely, but when it does, it’s almost never a mystery if you know where to look. A candidate can only be stuck in one of five specific places, and each one leaves a different, checkable trace — a missing score, an unopened invite, a locked booking, a knockout tag, or a verdict sitting in a queue nobody’s opened yet. This is the checklist, in the order to actually check it.

Stop one: did the sheet row ever trigger?

Start here, because it rules out the most common failure before you waste time looking anywhere else. If a candidate’s row was added to the connected Google Sheet but the trigger flow never picked it up, nothing downstream happened — no resume score, no invite, nothing. This is usually a sheet-connection problem, not a candidate-specific one: a malformed Sheet ID, a URL pasted into the wrong field format, or a sheet whose sharing permissions changed after the JD was first set up.

If you suspect this is where things broke, Why Your Google Sheet Connection Isn’t Triggering Interviews is the dedicated fix for this exact failure mode, including what the inline validation error looks like when a pasted URL doesn’t resolve to a real Sheet ID. One diagnostic shortcut: if only one candidate on an otherwise-working campaign is missing, this stage almost certainly isn’t the problem — a broken trigger connection fails for every new row, not selectively.

Stop two: did the resume actually get scored?

If the trigger fired but you don’t see a resume match score against the candidate’s name anywhere in the JD’s pipeline view, the enrichment step may have failed silently — most often because the resume file itself was unreadable (a scanned image instead of parseable text, a corrupted upload, a format the parser doesn’t handle) rather than because of anything wrong with the scoring logic itself.

This is worth checking early because it’s easy to misdiagnose as a scheduling problem: no score means no invite was ever eligible to go out, which looks identical from the outside to an invite that was sent and ignored. The difference matters for what you do next — one is a re-upload, the other is a follow-up email.

Stop three: was an invite sent but never booked?

If the resume score is there and cleared the threshold, the next thing to check is whether a self-schedule invite actually went out, and whether it was ever opened. A candidate can sit here indefinitely with nothing visibly “wrong” — no error, no lock, just an unopened email. This is the single most common reason a candidate looks stuck, and it’s the one place a HireQwik status alone won’t tell you whether the candidate is ignoring the invite or never received it (a bounced email, a typo in the sheet’s email column).

The fix here isn’t a technical one — it’s a manual follow-up, or in the interim, a nudge from the recruiter directly. This specific stop accounts for a meaningful share of every campaign’s overall drop-off, not just this one candidate’s case, which is why reminder cadence on the invite email matters more than almost anything else in the pipeline.

Stop four: is the booking locked?

If the candidate did book, then tried to move the slot more than three times within 72 hours of that first booking, the slot picker returns an HTTP 423 Locked response and the booking freezes — deliberately, so a drive with a real close date doesn’t have bookings drifting indefinitely. From the candidate’s side, this looks like the reschedule link simply stopped working. From HR’s side, it’s visible as a locked booking that needs a manual reset.

This is one of the more satisfying stops to diagnose, because the fix is a single, well-defined action rather than a guess: The 3-Reschedule Limit explains exactly what the lock looks like and how to clear it.

Stop five: the interview happened, but where’s the verdict?

If the candidate completed their call and you still can’t find a verdict, check the Needs Review tab specifically before assuming something’s broken. Not every completed interview auto-decides — a Go or On Hold verdict sits in the reviewed middle band by design, because neither evaluator produced a strong enough read in either direction for auto-decide to settle it without a human opinion. The Score/Decision filter on that tab lets you narrow straight to the band a candidate is likely sitting in, rather than scrolling the full queue looking for one name.

If a candidate genuinely isn’t showing up anywhere in the Needs Review queue despite a completed call, check whether the call ended on a phase-0 knockout — those end with a Reject tagged “knockout,” which is a different (and findable) status from a candidate still awaiting review. Inside the HireQwik Dashboard covers what each of these statuses actually looks like on screen.

A real case: the candidate who “disappeared” for six days

A recruiter running a mid-size fintech’s campus drive once flagged a candidate as missing entirely — no score, no invite, no verdict, six days after the application came in, on a campaign where every other candidate from the same batch had already been processed. Working the checklist above: stop one confirmed the trigger was firing fine for everyone else, which meant the sheet connection itself wasn’t the problem. Stop two turned up the actual cause — this one candidate’s resume had been uploaded as a scanned photo of a printed page rather than a text-based PDF, so the enrichment step had nothing to parse and silently produced no score.

Nothing about this was a HireQwik failure exactly — the enrichment step did exactly what it was supposed to do with an unreadable file, which is nothing — but it also wasn’t visible without knowing to check stop two specifically. The fix was a two-minute one: ask the candidate to resubmit a text-based file, which retriggered scoring and put them back on the same timeline as everyone else in their batch. The six-day gap wasn’t the pipeline being slow — it was six days of nobody looking at the one place the failure actually lived.

Two questions worth asking before you escalate

Is this one candidate, or a pattern? If a single candidate is stuck, work the five stops above — it’s almost always resolvable in minutes once you know where to look. If several candidates from the same campaign are stuck at the same stop, that’s a different problem: a sheet-connection issue, a resume format that a specific source system exports badly, or a threshold configured too aggressively. Fix the pattern, not each candidate individually.

Does “stuck” actually mean “waiting on a human”? Not every candidate sitting without a final verdict is stuck in the broken sense. A candidate in the Needs Review band is exactly where the process intends them to be — sitting until someone reviews the queue, not stalled because the pipeline failed to do something it should have. Confirming which of these two situations you’re actually looking at is often the fastest step in the whole checklist, and it’s usually the first thing worth ruling out before assuming something technical has gone wrong.

Why this happens more than it should

SHRM’s State of AI in HR 2026 research found that a share of organizations using automated hiring tools reported those tools overlooking or screening out qualified applicants — not because the tools were wrong on purpose, but because nobody was checking where the filtering actually happened. A candidate who’s “stuck” in HireQwik is almost never a mystery once you know the five stops above; the risk is a recruiter assuming the pipeline is a black box and giving up on the search after the first place they check comes back empty.

What tends to cause each stop, in practice

The five stops above aren’t equally likely. In practice, stop three — an unopened or ignored invite — accounts for the largest share of candidates that look stuck, simply because it’s the one stage that depends entirely on the candidate’s own attention rather than anything HireQwik controls. Stop one, a broken trigger, is rare once a campaign is running normally, but common in the first day or two after a JD is first connected, which is exactly when a Sheet ID gets pasted wrong or a sharing permission gets missed. Stop two, an unreadable resume, shows up disproportionately with candidates applying from mobile devices or through third-party job boards that convert file formats inconsistently on export. Stop four, the reschedule lock, is genuinely rare — most candidates who reschedule do it once — but it’s also the fastest to diagnose, since a 423 response is unambiguous. Stop five is less a failure than a reminder that “completed the interview” and “has a final verdict” aren’t the same status, and treating them as interchangeable is what makes a perfectly normal Needs Review candidate look like a bug.

Knowing this distribution changes where it’s worth spending prevention effort. A recruiter who wants to reduce how often candidates get stuck at all is better off tightening reminder cadence on the invite email (stop three) than worrying about trigger reliability (stop one), which mostly takes care of itself once a campaign’s first few days have gone smoothly.

The fast version of this checklist

For a candidate who genuinely can’t be found: check the trigger first (does the campaign work for other candidates, or is everything broken), then the resume score (present or absent), then the invite (sent, opened, booked), then the lock status (423 or not), then the Needs Review queue with the Score/Decision filter narrowed to their likely band. Five checks, in that order, and one of them will have the answer. What almost never happens is a candidate silently vanishing from all five — HireQwik doesn’t delete pipeline state, so “stuck” always means “sitting at one specific stop,” not “gone.”

For the full pipeline this checklist is built on, see From Resume Upload to Shortlist. And if the honest answer is that the candidate cleared everything and simply hasn’t been reviewed yet, What Happens After a Strong Go covers what the next step should actually look like once they do.

Questions about a specific campaign’s setup are also worth raising directly — book time with the team if a pattern like this keeps recurring across candidates rather than showing up once.