How a HireQwik Resume Upload Becomes a Shortlist

A HireQwik resume upload shortlist starts when the resumes are already files, not rows. I learned that the hard way in June 2026. A hiring team arrived with a folder of PDFs and no spreadsheet. Our only intake at the time wanted an Excel of Drive links. They would have spent a day uploading each file to Drive and typing a link column before anyone got a score. That was the wrong first step, so we shipped a second intake on 29 June 2026. You pick the job, drop the files, and HireQwik builds the shortlist.

This post is the map of that path. It stops at the shortlist on purpose. A later voice interview is a different screen, covered in from resume upload to the full pipeline. If you only have a folder, stay here.

What a HireQwik resume upload will take

On the Shortlist screen there are two ways to bring people in. One is an Excel whose columns include a resume link. The other is Upload resume files, which takes PDF and DOCX only. You choose one. The server refuses a job that tries to send both, because the two shapes do not share a column map.

The limits are plain. At most 500 files. At most 10 MB on any one file. At most 500 MB across the batch. A .doc from an old Word install, a photo, or a zip of resumes will not enter the job. Split a huge dump rather than hoping the cap will stretch.

Identical bytes are dropped inside that same upload. If someone saved Rahul.pdf and Rahul-copy.pdf and the bytes match, the second file never becomes a row. A rewritten resume for the same person, with different bytes, still becomes its own row. We do not pretend those are the same person at upload time.

A recruiter eye-tracking study often cited by HR Dive found reviewers glancing at a resume for about seven seconds. A folder of 200 files is not a seven-second problem. It is a morning. The upload exists so that morning is spent on the cut, not on renaming files into a sheet.

Why HireQwik asks for the job description first

The files do not score against a generic “good candidate” idea. You must pick a job description that already lives on the Jobs page. If you skip that pick, the job is not created. We removed the old habit of pasting a JD on this screen, because a pasted copy and the configured job drifted apart. The Jobs page is the single place the role is set up.

That choice decides the ruler. If the job has a structured resume scoring matrix, HireQwik uses that matrix and ignores a generic skills list. If it does not, the relevance-aware scorer runs, and a keyword check is only the last fallback. How HireQwik matches a resume to a job description is the longer account of that ruler. When a job has no rubric is the case you are in if you never built the matrix.

I would not upload a mixed folder “for all open roles” and sort it later. One job, one ruler, one shortlist. A second role gets a second run.

How HireQwik gets a name out of the file

There is no name column, because there is no sheet. Until the text is read, the row label is the filename with the extension stripped. Final_Resume_v7.pdf shows up as Final_Resume_v7, and the email is empty. That placeholder is honest. We do not invent a person from a filename.

After the text comes out, HireQwik writes the name and email it actually found. A campus CV that puts the email in the header will show that email. A file whose text is missing, or shorter than 50 characters, is marked failed to parse and is not scored. The filename stays. You can see the failure instead of a fake zero.

That is also why a scanned photo of a printed CV often dies here. The page looks fine to you. The text layer is empty. The same class of miss is described in why a candidate looks stuck. On this screen the status is on the row, not a silent hole in a sheet.

How HireQwik scores the uploaded resume

Each file is handled in batches of 50. The score that lands is on a 0 to 100 scale. Irrelevant history is not a bonus. A long CV with no skills and no experience that match this role is held to a low score. Years and a degree, counted with no check for the role, used to squash almost everyone into a band around 50 to 60 percent, which we wrote up in why resume scores cluster. The upload uses that same scoring stack. The intake changed. The ruler did not.

You will not get a shortlist that means “best writer of a filename.” You get a shortlist that means “cleared this job’s ruler.” If the ruler is thin, the shortlist will be thin in judgment too. Fix the job description before you blame the folder.

SHRM’s 2026 research found that 39 percent of organisations have put AI into HR work. A score is only useful if a person can see why it landed. On the review row, the reason sits next to the number. Open the person and you get the file plus that reason, not a naked percentage.

Where the HireQwik shortlist cut sits

When the run finishes, HireQwik applies one cut. The default is 60. Every completed score at or above 60 is marked shortlisted. Every completed score under 60 is not. Failed parses are not in that math. They never received a score.

You can move the cut later. Moving it re-marks the rows. It does not read the files again. I treat 60 as a starting line for a first look, not as a law. A campus role with a wide net may want a lower cut so you see more Maybes. A specialist role may want a higher cut so the table is short. The mechanics of moving that number are in how the HireQwik shortlist threshold works.

This cut is not the four bands on the Needs Review inbox. Those bands (Strong Go from 90, Go from 75, Maybe from 60, Reject below) belong to the sheet-triggered queue. Do not brief a hiring manager with inbox labels if you ran a file upload. Brief them with the score and whether the row is on this shortlist.

What you do with the HireQwik review table

The table can show only shortlisted rows, or only the ones that missed. You can sort by score, search a name, and export. Opening a row shows the resume and the reason for the score. You can also record your own approve or reject. That decision is yours. The automatic mark only says the score cleared the cut.

If the run is still going and the job was the wrong one, Stop screening halts it. Partial results stay so you can see who already scored. I would rather stop a bad ruler early than wait for 400 confident mistakes.

A practical order I use:

1. Confirm the job title on the run before you trust any score.
2. Open two high scores and two low scores and read the reason against the CV.
3. Then move the cut, or leave it, and export the shortlist you will actually call.

The review lives with the rest of the recruiter tools at the HireQwik HR dashboard.

What a HireQwik shortlist does not do

It does not interview anyone. It does not send a calendar invite. It does not apply the Strong Go or No Go labels from a voice call. Those steps come after a human decides the file cut is worth a conversation. From a resume shortlist into voice screening is that next handoff, and it is optional. Plenty of teams only wanted the ranked files.

It also does not replace a spreadsheet when you already have one. If every row already has a name, an email, and a Drive link, the Excel intake is the better door. The file door is for the folder. HireQwik direct resume upload versus a link sheet is the choice, written out. A drive-day version of the same folder problem is in a HireQwik campus resume shortlist. Files that die before a score are in resume files that never reach the shortlist.

The take I want a hiring manager to leave with: uploading is not the shortlist. The shortlist is the scores against one job, cut at a number you can move, with the failures left visible. If you cannot explain a high score from the reason on the row, do not send that name forward.

How I watch one HireQwik upload from the first file to the export

I stop treating the run as a black box the moment the button is pressed. A folder of 120 PDFs does not finish as one lump. HireQwik walks them in groups of 50, so the first group can already show scores while the third group is still being read. I keep that tab open. If the earliest rows are all filenames with no emails, I do not wait for file 120. The text layer is wrong, and more waiting only adds more of the same miss.

Here is the walk I use on a single person, call her the first completed row, so a new recruiter can see the whole path without a lecture.

She starts as a filename. There is no typed email, because nobody built a column. After the text is pulled out, the row label becomes the name printed on the CV and the address printed with it. I open that row before I look at anyone else. If the name on the PDF and the name on the row disagree, I do not trust the rest of the batch yet. A heading parsed as a person is a warning about the file, not about her skill.

Her number is then 0 to 100 against the job I picked, not against a general idea of a strong graduate. I read the reason in the same glance as the PDF. The reason should mention skills or work the job actually asked for. If it praises a long list of tools the JD never named, the ruler is loose and I fix the job before I export. Scoring does not get a second, kinder pass for people I like. The file is judged once. My judgment is the decision I type afterward, separate from the automatic mark.

While later groups are still moving, I can already sort the finished rows. I do not export yet. An export halfway through is a lie, because the people still in the queue are absent, and absence looks like rejection. I wait until the job says it is finished, or I hit Stop screening because the job title on the run is the wrong one. Stop leaves the rows already judged. It does not pretend the unfinished files were scored. I would rather brief a manager on a partial, labeled partial than on a pretty list that hid a bad ruler.

Only then do I filter to the marked rows and download that set. The note I attach has four lines: which job, which cut, how many files I dropped in, how many never received a number. That fourth line is the one managers skip, and it is the one candidates ask about. A missing number is not a low score. Say so in the note or you will spend the afternoon explaining a rejection that never happened.

I also keep the original folder for a week. HireQwik’s review shows the file, but the folder is what the sender will recognize when you ask for a replacement. Matching a failed row back to Final_v7 is much faster if you still have the zip. Delete it after the panel has the list, not before.

None of this is a second product. It is the same Shortlist screen, watched in order: files in, names filled from the text, a number against one job, a mark at the cut, an export that admits who never got a number. Teams that skip the watching step get a spreadsheet of scores and a fight. Teams that watch the first group usually catch the wrong job in ten minutes.

If you want this run on your own folder, book a walkthrough and bring the job description plus a small batch of real files. A sample of twenty tells you more than a slide.