When an HR team tells you they have more than 6,000 applicants across five job posts, the obvious solution is AI scoring. Feed in the CVs, let the software rank everyone from best to worst, and hand recruiters the top fifty.

Here's why, and what any HR team with too many applicants can take from it.

The real problem wasn't too many candidates

At first glance, 6,000 applicants looks like a volume problem. It isn't. It's a format problem.

A sortable spreadsheet answers all of them in seconds.

Why we didn't build a scoring algorithm

We had three reasons, and they apply to most high-volume hiring.

1. Scoring needs clean data, and the data was not clean

A scoring tool is only as good as what it reads. Here's a small example from this project. Many applicants had added certifications to their names: PMP, CSM, PSM, MBA and others. The team asked for clean names only.

2. The people who know the client should make the call

Atlin's recruiters know what their clients actually value: the kind of experience that transfers, which certifications matter in which markets, and what a strong candidate looks like beyond the keywords.

3. Ranking people is where the legal risk lives

This is the reason more HR teams should be talking about. A tool that organises applicants is admin software.

New York City requires independent bias audits for automated hiring tools. The EU AI Act treats AI used in candidate selection as high-risk. And in the US, the Mobley v.

What it saved, in plain numbers

The fair comparison isn't "automation versus nothing." It's automation versus the recruitment coordinator who would otherwise do it. You can run the same sum with your own volumes in an automation savings calculator.

If handling one applicant by hand (opening the application, filing the CV, copying the details into a spreadsheet, checking them against the survey list) takes about 3 minutes, then 6,000 applicants is roughly 300 hours.

The bigger saving isn't the wages. It's that the coordinator spends those weeks talking to candidates instead of typing them into a spreadsheet.

Lessons for any team drowning in applicants

Start with structure, not intelligence. Before buying anything labelled "AI screening," ask whether your team simply needs every applicant in one sortable list. Often that alone solves most of the problem.

Show a real spreadsheet early. We delivered the first role on day two, and the team's feedback shaped every role after it.

Is a sorting pipeline right for you?

It's a strong fit if single roles attract hundreds of applicants, your team copies candidate details into spreadsheets by hand, or you need to combine applicant data with a survey or assessment. That is the kind of workflow a custom automation development project can take on, starting with a single role.