Step 1 (days 1–3): choose one role and write the scorecard with its hiring manager
Choose an inbound-heavy role with a clear profile, such as one where applications arrive in volume and hard requirements exist. Sit with its hiring manager and write the scorecard: the hard requirements that must be evidenced, the weighted criteria, and for each what evidence in an application would satisfy it. Record the baseline from the last two rounds: applications received, time from close to shortlist, share of interviewees who met the hard requirements, and candidate response times.
The scorecard is the pilot's real product. A Recruiter Digital FTE applies it consistently with evidence; it cannot make a vague scorecard precise.
Step 2 (days 3–5): connect the ATS and write the remit
Connect the applicant tracking system with a dedicated account scoped to the role's pipeline, with permission to read applications, draft communication, and schedule into connected calendars, and no permission to change a candidate's status. Write the remit: owns screening, shortlists, drafted communication, and scheduling; escalates any candidate the scorecard captures poorly and anything about a candidate's circumstances; never decides or communicates a rejection or offer.
The last item is enforced by the platform's never-automated list regardless; writing it in the remit makes the team's expectation explicit.
Step 3 (week 2): review shortlists with the evidence per score
As applications arrive, shortlists appear in Approvals with each candidate's scores and, for every score, the passage in the application it rests on, plus flags for candidates who need a human conversation before ranking. The hiring manager reviews by criterion: agree, or send back a specific score with a reason. Drafted candidate communication waits for approval; scheduling proposals wait too.
This week establishes the record. Read closely, send back precisely, and keep everything gated.
- Hard requirements: evidenced, not evidenced, or flagged for verification
- Weighted criteria: score and the passage behind it
- Flags: candidates the scorecard captures poorly
- Communication and scheduling: drafted, awaiting approval
Step 4 (week 3): tune the scorecard and run the first fairness check
Read the send-back reasons grouped by criterion. A criterion corrected repeatedly in the same direction is a criterion written wrong; rewrite it with the hiring manager and version the scorecard with the date and reason. Then run the first fairness check: compare shortlist outcomes across the groups you monitor and, where differences appear, open the scorecards to see which criteria drove them. Because every score shows its evidence, the audit is a review of criteria rather than of a black box.
Record the check and any change. That record is what a fair-process review will ask for, and it is more defensible than any claim of neutrality.
Step 5 (week 4): review and the scheduling decision
Hold the first monthly review from the FTE's report: time to shortlist, share of interviewees meeting hard requirements, approval rate per criterion with reasons, usage. Decide whether interview scheduling into connected calendars has earned autonomy from a clean record; it is the one class that changes no outcome and so the one class that can. Rejections and offers are not on the ladder, and communication about a candidate's status stays gated.
Log the decision. Then plan the second role, whose scorecard will be written faster because the team now knows what a scorecard that works looks like.
What does done look like at day 30?
Done is a role whose applicants were screened against a written, versioned scorecard with evidence per score, shortlists that reached the hiring manager within a target window, a first fairness check recorded with its findings, every rejection and offer made by a named person, and a scheduling class either promoted from a clean record or deliberately kept gated. The baseline comparison is read at the second review.
The hiring manager spent minutes per shortlist and an hour on the fairness check, and has a screening process they can explain to a candidate, a regulator, or a new colleague in a sentence: these criteria, applied the same way to everyone, with the evidence shown, and a person deciding.
What should you refuse to do in the first 30 days?
Refuse to let the scorecard drift by editing shortlists instead of criteria; edits leave no reason and fix nothing. Refuse to promote anything but scheduling, however clean the record, because the other classes touch a person's outcome. Refuse to skip the fairness check because the numbers look fine; the check is what makes the numbers defensible. And refuse to start a second role before the first review, because the scorecard lessons from the first are what make the second fast.
These refusals are what a fair, fast screening process is made of, and they cost nothing but discipline.