Skip to main content
GullySystem
AI for HR · Solutions

Applications arrive in a suggested reading order.

AI Candidate Screening reads resumes and assessment scores for one role and suggests a reading order, with matching and missing items listed beside each name. This is decision support for a person. It flags and ranks, a qualified recruiter decides, and it makes no claim to remove bias or judge anyone’s suitability.

What it is

Reading all of them takes care. A single vacancy for an accounts executive can bring hundreds of resumes in different layouts, and the recruiter reads the first few carefully and the rest quickly. Order of reading becomes the decision.

GullySystem sets up a screening step that lists each applicant’s matching skills, missing requirements and date gaps, so the recruiter knows where to look first. Nobody is hidden from the recruiter. Every resume stays one click away.

Where it goes wrong today

The problems this is meant to ease.

Order of reading becomes the decision

Hundreds of resumes arrive for one vacancy, and the first few get care while the rest get seconds. Strong applicants who applied late are never properly read.

Different recruiters apply different standards

One recruiter treats a missing certificate as a stop, and another asks about it. Candidates are handled by habit, and nobody can explain why two similar profiles went different ways.

Information missing and nobody asks

A resume leaves out dates or a required licence, and the application sits in a folder. The gap is found at the interview, or never.

Resume layouts hide good candidates

Tables, images and unusual formats confuse keyword searches. Someone with the right skills is passed over because the file would not read.

Comparison done from memory

At the shortlist meeting, recruiters recall candidates from impressions. Notes are patchy, and the hiring manager cannot see on what basis two people were compared.

Unexamined screening shortcuts

Teams rely on school names, past employers or years of experience that the role does not truly need. Nobody reviews whether the shortcuts exclude people for no job-related reason.

Inside the module

What AI candidate screening does.

Requirements written as a checklist

The hiring manager and recruiter turn the job description into a short checklist of must-haves and nice-to-haves that both sign off before applications are read. Every item must relate to the job, and a weak checklist gives a weak ranking.

Qualifications and experience extracted

Employers, titles, dates, tools and qualifications are lifted from each resume and cover note into fields you can sort and filter. Unreadable files are listed apart, never silently dropped.

Matching against stated criteria

Each applicant is set against the checklist, item by item. Next to every name sit the items found, the items not found, and the resume lines each came from, so overruling takes one click.

Missing information chased

When a resume lacks a date, a licence or an answer to a required question, the gap is listed. A follow-up message is drafted for the recruiter to approve and send, and replies are filed on the record.

Side-by-side comparison

Shortlisted candidates are laid out in one table on the same checklist items, with assessment scores shown beside the resume reading. The recruiter sees what was measured, and one score is never the answer.

No verdicts on character

Nothing here infers personality, motivation, age, health or fit from photos, writing style or speech. It reports what the documents state against the checklist and stops there.

Gaps flagged, not judged

A break in dates, a short stay or a missing qualification is marked as a question for the interviewer. It is a prompt to ask openly, never a reason to set an application aside.

Checks on fairness and access

Before go-live, and again every few months, the HR manager compares rankings with the shortlists recruiters chose and looks for unexplained gaps between groups. Accessible forms and alternative formats are tested with real users.

Where it gets used

What this looks like in a working week.

A recruitment agency

A client vacancy brings several hundred applications. Each is read against the agreed checklist, and the recruiter sees reasons beside every name before presenting a shortlist.

A hospital or clinic

A ward needs nurses with a valid registration and a stated specialism. Missing registration details are listed, and a follow-up message is drafted for the HR officer to send.

A software company

Applicants submit resumes and a short coding exercise. Resume reading and exercise scores appear side by side, and the engineering manager checks the reasoning before anyone is invited.

A security agency

Guards are hired in large batches. Licence dates, shift availability and previous posting history are extracted, and a supervisor reads each flagged gap before a call is made.

A hotel

Front desk and kitchen roles bring applications in several languages. Relevant experience is listed in one table, and roles judged mainly by a trial shift go straight to a practical test.

Who touches it

The people who open this every day.

Recruiter

Reads in the suggested order. Opens the resume behind every line, adds notes, and decides who moves forward, who waits and who is declined, in the recruiter’s own words.

Hiring manager

Agrees the checklist first. Fixes the must-haves before applications arrive, and later looks at the shortlist to say whether the people on it match the job as they see it.

HR manager

Checks how it is used. Reviews the checklists and the override notes, and takes any question about hiring practice to a qualified employment adviser rather than to the software.

Interviewer

Reads the flagged questions. Asks the prepared prompts openly during the interview, which gives each candidate a chance to explain a gap before anyone forms a view of it.

How it is set up

From first conversation to daily use.

  1. Discovery of how screening happens now

    Watch a recruiter read a batch of applications. Record which items they look for, which shortcuts they use, and how decisions and reasons are written down in the tracking system.

  2. Data readiness for resumes and checklists

    Collect past resumes with the outcomes that people chose, write a checklist per role with the hiring manager, and remove items that do not relate to the work.

  3. Trial on a filled vacancy

    Run the reading on one role that has already been filled. Compare the suggested order with the real shortlist, and list every difference with the recruiter.

  4. Checking the output with recruiters and managers

    Recruiters and the hiring manager read reasons beside the ranked names, correct wrong readings and rewrite checklist items that confused the matching. Accessibility testers try the forms and messages.

  5. Going live with a person at every step

    Switch on the reading order for one live vacancy. A recruiter opens every resume that matters, and overrides are recorded with a short note from the first day.

  6. Reviewing afterwards

    The HR manager reads overrides, missing-item counts and fairness comparisons on a regular schedule, takes concerns to an employment adviser, and rewrites checklists that keep producing disagreement.

Measuring it

How you would know it is working.

Record where you stand before it goes live, so the change is measured against your own figures.

  • Processing time is the recruiter hours spent reading each batch of applications, estimated from the last few vacancies filled before go-live.
  • Screening consistency compares how often two recruiters reading the same sample of applications reach the same shortlist, tested before and after.
  • Recruiter review quality is the share of ranked names where the recruiter agreed with the stated reasons, taken from override notes after the trial.
  • Missing-information rate counts applications that arrive without a required item, compared with a hand count over recent vacancies.
  • Candidate follow-up completion records how many gaps were queried, using the messages logged in the tracking system.
Used responsibly

Where a person stays in charge.

Decision support, not a decision

It assists a recruitment team in reviewing job-related information. It does not decide who deserves employment, and every outcome, including declining, is chosen and recorded by a person.

No claim to remove bias

Ranking by a checklist can repeat the habits behind the checklist. No promise of fairness is made, so the HR manager tests results for discrimination risks and reviews overrides on a regular schedule.

Job-related information only

Only items on the approved checklist are used. Personal traits, photos, social media and anything unrelated to the work are not read, and applicants are told what is read.

Privacy and local rules

Applicant data stays in your tracking system or the database set up for you, with access limited to named people. Countries differ on automated screening and retention, so settle both with a lawyer you appoint. It only organises data and offers no legal or employment advice.

Accessibility for applicants

Forms, messages and assessments are tested with assistive tools, and applicants can ask for another format or a person to read their application.

How it fits

It is one part of a system, not an island.

A module earns its place by what it passes to the next one. These are the connections we set up most often.

  • Candidate tracking
  • Resume management
  • Interview scheduling
  • Skills assessment tools you already use
  • HR analytics and reporting
Common questions

Questions owners ask about AI candidate screening.

Does it decide who gets hired?

No, a person decides. It flags and ranks, a qualified recruiter decides every outcome, and nothing it shows is legal, employment or hiring advice or a verdict on any person.

Can a recruiter disagree with a rank?

Yes, at any point. Any placement can be overruled with one click and a short note, and those notes are kept for the HR manager to read when the checklist is reviewed.

What does it need before it works?

A written checklist for each role, resumes in readable files, and a recruiter who will actually read the reasons behind every rank. Without that, it becomes a quiet filter. That is not its purpose.

Where are applicants’ details kept?

In your own applicant tracking system or the database we set up for you, with access limited to the people you name. Each country sets its own rules on holding applicant data. Retention periods and consent wording come from a lawyer you appoint.

When should we skip it?

When you receive few applications per role, reading every resume yourself is better and simpler, and you will notice things no ranking can. It also helps little for roles where good work is judged mainly in person or on a trial, such as a chef or a site supervisor.

Will it remove bias from hiring?

No. Any scoring method can repeat old habits. The goal here is a visible checklist and visible reasons, so people can question both, and a regular test for unexplained gaps between groups.

How accurate is the reading of resumes?

It makes mistakes, especially with unusual layouts, scanned files and non-standard titles. Anything it cannot read is listed apart, and the original resume is always one click away.

What does it connect to?

Your applicant tracking system, resume store, interview calendar and any skills test you already use. Each product decides how much it lets outside software do, so we scope the connection after seeing yours.

Talk to us

Tell us how you handle AI candidate screening today.

A spreadsheet, a register or another system: say which, and we will tell you plainly what is worth changing.

  • No obligation
  • A reply within one business day
  • Your details stay private

Start with your requirements

Seven short steps that tell us what you need. An engineer reads them before getting in touch.

Share your requirements