Why Don't Companies Give Reasons for Rejections

Imagine grinding for hours on LeetCode, practicing behaviorals, staring at your resume for hours, and then getting rejected.

It stings, but you understand that it happens. That's not the worst part.

The worst part about getting rejected from a company is not knowing why you didn't get the position.

Was it because I was unqualified?

Did I mess up the second coding question?

Did I yap for too long about the wrong answer?

Or was it because I was so tuff and overqualified that they were intimidated by me? Probably not.

If companies gave reasons for rejections, like the position was already filled or you didn't have the required skills, it would make life 10x easier.

Instead, you apply to a job and get the default message:

"Thanks for applying/interviewing. Unfortunately, blah blah blah, we chose the CEO's son to join our frontend team."

But if you got feedback on what exactly went wrong, you could maybe apply again to the same company and work on those skills. You could also understand that maybe you didn't do anything wrong, but it was just a competitive role or they wanted more years of experience.

The default rejection email doesn't give you a path forward. You don't know what to fix.

Should I work on:

The interview and application process has so many moving parts that a rejection with no feedback doesn't benefit the applicant or the employer. The applicant can't move forward, and the company isn't setting a clear standard for what they want applicants to meet.

The Technical Reality

Looking at it from a technical standpoint, how hard would it be to actually implement this feature? And more importantly, how do companies avoid PR nightmares?

My qualification on this topic is that I designed an automated system for a technical company that attempted to automate part of the hiring process through my AI/ML co-op at Quantum Business Advisory. The hardest part wasn't building the model, but rather the quality of the data.

If a company uses AI to give feedback to candidates before the interview stage, meaning they weren't qualified to get an interview because stronger candidates were found or the company believed they lacked certain skills, one wrongly generated feedback sentence could hurt the company.

Let's say a company builds an AI feedback generator where they genuinely don't have to do anything. The model detects why the candidate is unfit and says something like:

"The candidate is not qualified because he doesn't have skills in databases or Python and only has 3 years of experience."

But the candidate actually has a project where he used PostgreSQL and MongoDB to store user information, track records, and update data migrations through Alembic. Not only that, he has 6 years of experience, but couldn't fit all of it on the resume and instead highlighted the key experiences he had.

The AI rejects him.

AI looks at everything through a black-and-white lens.

Does this person have 3 years of experience on their resume? No, reject.

Does this person work with Kafka? No, reject.

The reason why this is stupid is because adding a bullet point about Kafka, GitHub Actions, or AWS:

a) doesn't mean you actually have those skills, it might just mean you tailored your resume

b) means you listed those skills, but doesn't mean you're proficient in them

AI doesn't understand the human element, and that's why my model kept failing. Using BERT wasn't the issue. Was it the data? Maybe. But it was also the fact that resumes can't determine the potential someone has (Megumi JJK reference), just by judging a resume.

And unfortunately, in this scenario, the solution isn't full automation.

We get into a habit of trying to automate every single moment in our lives: sending emails, writing posts, writing code for us. However, one wrongly generated feedback sentence could genuinely be worse than the default rejection message.

So What's the Solution?

I'm not sure yet, but I believe one option could be recruiters scanning a resume, highlighting the specific issue they noticed, and then asking AI to further explain the reason. That could work.

Or when someone gets rejected after an interview, since AI note takers are everywhere now, companies could generate a summary based on specific guidelines for what they wanted in the candidate. That way, feedback is given and time is saved.

I understand that some people may not want AI feedback, but recruiters have thousands of applications. There's no way they can take the time to go through every candidate and give meaningful feedback.

So the real question is: Is some feedback better than no feedback?