The AI application flood: what the data says, and a fair sift

Updated 26 October 2026

The short answer

Workday’s worldwide data shows more applicants for each filled job, and most job seekers in its survey say they use AI in their search. Some UK employers report ‘excessive’ use of generative AI in applications, and more employers saw a rise in unsuitable applicants than in suitable ones. When polish is cheap, sift on what is harder to fake: answers to the same short questions for every candidate, tied to the job’s must-haves and scored against a guide you agreed in advance, with a person deciding who goes forward.

Why it is worth fixing

The CIPD’s latest Resourcing and talent planning report, based on UK employers, says a sizeable minority of organisations that tried to recruit saw ‘excessive’ use of generative AI in applications. Many more employers reported a rise in applications from unsuitable candidates than from suitable ones. Only a minority try to monitor AI use, and most of those have rejected candidates for it. These are employers’ own reports, and the exact figures are on the CIPD’s page.

Workday’s Global Workforce Report, which draws on Workday’s own customer data and surveys from around the world rather than the UK alone, found the typical number of applicants for each filled job rose over the year, while the time to fill a role held steady. Workday’s reading is that more applications don’t necessarily mean more qualified candidates. The large majority of job seekers in its survey said they used AI in their search.

Neither report shows that an application written with AI comes from a weaker candidate. Our view: the answer is a better sift, not a hunt for AI.

Step by step

  1. Write the must-haves before you open an application

    List the few things the job truly needs, such as a licence, a qualification, availability or a skill shown at work, and keep them apart from the nice-to-haves. If you can’t say how you would check a must-have, it isn’t ready to sift on.

  2. Ask every candidate the same short questions

    Add a few questions to the application, each tied to a must-have, such as “Tell us about a time you…”, with a word limit. Everyone answers the same ones. A polished cover letter is easy to produce now; a specific answer about the candidate’s own work is harder to fake, and it gives you like-for-like answers to compare.

  3. Score against a guide agreed in advance

    For each question, write down what a strong, an adequate and a weak answer looks like before the answers arrive. Score every candidate against the same guide, with two people for roles that matter. Agreeing the guide first is what makes the sift fair, and what lets you explain it.

  4. Let software do the admin, and a person decide

    Software can take on the admin: acknowledging each application, collecting the answers in one place, flagging missing information, and laying answers side by side for a reviewer. It should not score, rank or turn anyone down. A person reads the answers, decides who goes forward, and can say why.

  5. Sift for the job, not for AI use

    Many good candidates now use AI to tidy their writing. Rather than trying to catch it, design the sift so it matters less: questions that need real examples, and a short follow-up conversation in which candidates talk through what they wrote. If you do set a rule on AI use, put it in the advert so everyone knows it.

  6. Tell candidates how you use AI, and offer adjustments

    Add a plain line to the advert and the acknowledgement, and adapt it so it is true for you: “We use software to organise applications. A member of our team reads every answer and makes every decision. If you need any adjustment to our process, just tell us.” Employers must make reasonable adjustments for disabled people and people with health conditions, and gov.uk lists changes to the recruitment process among them. The government’s Responsible AI in Recruitment guide explains how AI tools can create new barriers.

Checklist

When it is worth getting help

For a handful of roles, a shared form and a scoring sheet will do. It is worth getting help when applications arrive through several job boards and inboxes, when acknowledging, collecting and chasing is eating into your consultants’ time, or when you want a clear record of how each decision was made.

Rec Int. builds automation for UK recruitment agencies that takes repetitive admin off the team, around the systems they already run.

Questions people ask

Should we reject applications written with AI?

We wouldn’t make that a rule on its own. Many strong candidates use AI to tidy their writing. Sift on the answers to your must-have questions instead, and if you set a policy on AI use, say so in the advert.

Can software do the sift for us?

We would keep the sift with a person. Software can do the admin around it: collecting answers, flagging gaps and laying answers side by side. A person should read the answers and decide who goes forward.

Sources

If you would like to talk through the repetitive admin around your applications, such as acknowledging and chasing, book a discovery call.