ReviewDesk
← Back

ReviewDesk guide · By the Ingenious AI team · Published 14 November 2026 · Reviewed 14 November 2026

What must you tell job applicants when a hiring algorithm screens them?

Generate a free adverse-outcome letter — no sign-up Build a free pre-use notice See pricing

The short answer. Tell them before the tool runs that it will be used, what it screens for, and what role it plays in who gets interviewed — then, if the outcome goes against them, explain the principal reasons and offer a person's review. Illinois requires employment notice now; Colorado and California add their machinery from 1 January 2027. ReviewDesk holds the whole sequence in one hiring workflow: hosted notice, letter generator, review inbox and audit trail.

What is hiring-screening notices?

A hiring-screening notice is the pre-use disclosure owed to applicants when automated tools score, rank, filter or assess them for recruitment, hiring or promotion. In ReviewDesk's regimes it is the employment face of the ADMT duties: Illinois HB 3773 notice (already in force), Colorado SB 26-189 notice and 30-day explanation (from 1 January 2027), and California's ADMT notice with opt-out or appeal (compliance from 1 January 2027).

Where in the hiring process does the duty bite?

At every stage where software exercises discretion about a person: application sifting, knockout-question filtering, recorded-interview grading, assessment scoring, shortlist ranking, even interview scheduling when priority quietly follows a score. The duty does not wait for the final hiring decision — an applicant eliminated by a filter at stage one was decided about, and the notice must precede that filter. Map your funnel honestly and the notice points choose themselves: job posting, application start, assessment invitation — every door through which a candidate passes before a consequential filter.

The concealed filter is the dangerous one. Applicant tracking systems ship with scoring features that hiring teams enable without naming them; a 'match percentage' displayed to recruiters is a ranking with legal character when nobody below fifty per cent is ever viewed. The inventory method from the Illinois guide — walk the funnel, name every score — is the only reliable way to find them before an applicant does.

What does the applicant need to understand from the notice?

Four things, in candidate language. A tool is involved. What it examines — the factor groups, such as eligibility, experience markers, assessment performance. What it decides or influences — who is filtered, who is ranked, who is seen first. And what the applicant can do: correct their information, request a person to review, and in California opt out or appeal. Write it where candidates actually read — the application flow, not a careers-page footnote — and in the vocabulary of the job advert, not of the vendor's white paper.

Candidates judge fairness by procedural signals long before outcomes arrive. A clear notice measurably changes how a rejection is received, because the applicant who knew a tool was involved, and why, reads the outcome letter as the second half of a conversation. The applicant who discovers the tool afterwards reads everything — including your honest parts — as concealment. That asymmetry is the commercial case for notice, beyond the legal one.

What do adverse hiring explanations look like?

Specific to the stage and the factors. 'Not interviewed because the recorded assessment scored communication below the benchmark for client-facing roles, and shift availability did not cover two required patterns' is an explanation. 'We received many strong applications' is a courtesy, and courtesies do not discharge duties where explanation duties apply — Colorado's 30-day window covers adverse consequential decisions in employment among its areas. Generate the letter from the screening record: the stage reached, the factors recorded, the reasons entered — the pattern in the principal-reasons guide.

Volume is the operational challenge. Hiring funnels produce adverse outcomes by the hundred, so letters must be generated, not drafted — but generation without per-applicant reasons produces the 'many strong applications' failure at scale. The generator's structure solves the tension: the template carries the jurisdiction note and the review path, while the reasons are entered per applicant from their screening record. Batch the mechanics, never the facts.

How should review work for rejected applicants?

With a reviewer who can actually reopen the funnel. A meaningful review for a filtered applicant means a recruiter or hiring manager re-examining the full application — not the score — with authority to advance the candidate. The standard from the human-review checklist applies unchanged: full file, competent reviewer, real time, real authority. The most common genuine errors reviews find are parsing failures — qualifications misread, experience undercounted — which is why the correction route deserves equal billing with review in every outcome letter.

Respect the calendar of hiring. A review answered after the role is filled is a courtesy apology, not a remedy. Build the inbox rhythm around live requisitions: requests triaged daily while a search is open, reviewers assigned from the hiring team, and where the original role has closed, an honest outcome — considered for the next matching opening, with the record kept — beats silence. ReviewDesk's SLA clocks and case assignment were built for exactly that tempo.

What about promotion and internal mobility?

The regimes' employment language reaches promotion as well as hiring — Illinois says so expressly — and internal candidates are the audience most organisations forget. An employee shortlisted out of a promotion round by a performance-scoring tool is an affected person with the same notice interest as an external applicant, and with a longer memory. Internal notices belong in the internal mobility flow, written for employees, and review routes must handle the sensitivity of an employee challenging a process their own managers run.

Internal cases also surface the scope question sharply: performance analytics that 'inform' promotion discussions can materially influence them, and the scope worksheet should record the determination for internal decision types separately from external hiring. The test method is in the scope-test guide. One workflow per decision type keeps the notices honest — the promotion notice is not the hiring notice with a different header.

How do agencies and employers split the duties?

Staffing arrangements blur ownership: the agency screens, the employer hires, the applicant experiences one process. The duties follow the decision about the person, and both parties' names may belong in the notice — the tool's user and the deciding employer — with the review route landing somewhere a person actually answers. Contract for it expressly: whose notice, whose inbox, whose letterhead on explanations, and how records are shared when the applicant asks. The reseller and agency operating model is detailed in the HR-tech guide and the multi-business pattern in the agency guide.

What never works is the gap: agency assumes employer notifies, employer assumes agency reviews, applicant discovers the tool from a third party. If you buy screening as a service, the notice is still about your decision — put your name on it, run the inbox where your team can see it, and let the contract argue about cost allocation afterwards.

Your five actions for hiring-screening notices

  1. Walk the whole funnel — sift, filters, assessments, ranking, scheduling — and name every score that affects a candidate.
  2. Place the notice before the first consequential filter: posting, application flow and assessment invitation.
  3. Generate outcome letters from the screening record, with the stage reached and per-applicant principal reasons.
  4. Run reviews at hiring tempo: daily triage while a search is live, reviewers with authority to advance a candidate.
  5. Cover promotion and internal mobility as their own workflows — never as an afterthought to external hiring.

Worked example: a retailer's seasonal hiring, screened openly

A fictional retailer — Canyon Outfitters — hires 240 seasonal staff from 1,900 applications using a screening tool that checks eligibility, availability patterns and a short recorded assessment, then ranks candidates for interview. Its notice sits in the application flow: the tool, the three factor groups, the ranking's role, the correction and review routes. Outcome letters for the 1,660 not interviewed are generated from each applicant's screening record — stage reached, factors, reasons — inside a week of each decision, not in a January clear-out.

Reviews arrive from 74 applicants. The review bench — store managers with authority to interview — finds the patterns this guide predicts: nine availability records misread from a form redesign, six assessments scored on incomplete uploads, and a long tail of correct decisions explained better the second time. Twenty-one candidates are interviewed after review; eleven are hired. The funnel took the same time as the previous year's manual chaos, and produced something the manual year never did: a record that could answer any applicant's 'why' in minutes.

Frequently asked questions

When must applicants be told about a hiring tool?

Before it is used on them — Illinois employment notice duties are already in force, and Colorado and California duties apply from 1 January 2027. The notice belongs in the application flow before the first consequential filter, not in a rejection letter.

Do knockout questions and match scores count?

If they filter, rank or materially influence who progresses, treat them as in scope: a match percentage that decides who is viewed is a ranking. Record the scope determination per stage with the worksheet method.

What should a rejection explanation contain?

The stage reached, the principal reasons drawn from the applicant's screening record, the correction route for wrong information, and the human-review offer. Generate it from the record — courtesies about strong fields do not explain a decision.

Who should review a rejected applicant's case?

A recruiter or hiring manager with the full application, time to judge it, and authority to advance the candidate. Reviews must run at hiring tempo — an answer after the role closes is an apology, not a remedy.

Is this guide legal advice?

No. ReviewDesk is templates and workflow tooling, not legal advice. Duties are summarised from the regimes in our research; consult counsel for your situation.

Keep reading: How do ADMT notice and review duties apply to automated credit decisions? · What must landlords and screening services disclose when screening is automated? · How do notice and human-review duties apply to automated insurance underwriting and claims? · What notice and review duties arise when software helps make healthcare decisions about patients? · What must schools and universities disclose when software screens applicants or students? · All ReviewDesk guides · ReviewDesk home · Try the free desk · Free letter generator · Free notice builder · Pricing

ReviewDesk is built by Ingenious AI. Templates and workflow tools — not legal advice.