The application form is costing you candidates
Every extra field has a price. Most teams have never measured it.

Open your careers site on a phone, on mobile data, and apply to your own job. Time it. Most people who try this discover a fifteen minute process with an account creation step, a resume upload, and then a form asking for everything on the resume again.
Applicants abandon that. You never see them, so it never appears in a report.
Measure before you argue
Your ATS can tell you how many people started an application and how many finished. If it cannot, put analytics on the page. The number is usually between forty and seventy percent abandonment. Once you have it, the conversation with whoever owns the system stops being about opinion.
Cut ruthlessly
Ask only what you need to make a screening decision. Name, contact, resume or profile, work authorisation where legally relevant, and at most two role specific questions.
Everything else can wait until someone is in process. Address, education dates, previous salary where it is still legal to ask, emergency contact, diversity questions that could be voluntary and post application. None of that helps you decide, and each field costs you people.
Kill mandatory account creation
Requiring a password before someone has applied is the single largest drop off point in most systems. If your ATS forces it, that is a procurement conversation, and it is worth having.
Let people apply with what they have
Resume upload plus parsing, profile import, or a short form. Give options. Someone applying from a phone at lunch will not retype their history.
Then test it
Run the change on a subset of roles for a month and compare completion. Two things usually happen. Applications go up, sometimes sharply. And screening quality does not fall, because the fields you removed were never used to screen.
The teams that resist this argue that friction filters for motivation. It does not. It filters for people who are unemployed and have time, which is not the same population as people who are good.