Why Graduate Hiring Funnels Break at the Resume Screen — and What’s Replacing It in 2026

The graduate hiring funnel in India isn’t broken at the top. Sourcing works. Naukri, LinkedIn, and on-campus drives bring in applications at volumes most TA teams have never seen — campus seasons that pull in tens of thousands of applications per company are now routine, not exceptional. The funnel breaks one stage later, at the resume screen, and the damage compounds before anyone notices. By the time a hiring committee asks “where did our pipeline go?”, the answer is uncomfortable: we rejected most of it without anyone reading it.

The math nobody wants to do

Industry data is blunt about it: roughly 75% of resumes never reach a human reviewer (source), and recruiters today manage about 93% more applications than they did in 2021 without proportional headcount growth (source). For tier-1 IT services and BFSI firms running campus drives in 2026, the typical funnel for a 30-college season has the biggest drop-off between “applications received” and “resumes a recruiter actually reads.” That’s the leak. It’s invisible because nobody documents the rejected resumes — they sit in the ATS, marked “low keyword match,” and the funnel report shows healthy conversion rates from the stages where humans are looking.

The leak happens because every ATS in production was built for a different problem. ATS platforms are pipeline trackers. They were designed to log candidate state, route approvals, and produce quarterly reports. Resume screening got bolted on as a feature — usually a keyword-match algorithm with a confidence score — because customers asked for it, not because anyone built it carefully. The result is a screen that rejects qualified freshers because their CV doesn’t say “Java” but says “JVM,” “Spring Boot,” and the name of an actual project.

Why the bottleneck moved

Three things changed in the last 24 months and none of them got the attention they deserved.

First, application volumes per role roughly doubled. The typical campus TA team in India is operating with roughly the same headcount it had three years ago, against a candidate volume that has not stopped climbing. You cannot read every CV by hand at this scale; nobody can.

Second, AI-generated resumes broke ATS keyword matching. When a candidate uses a chatbot to rewrite their CV against the job description, the keyword-match algorithm gives them a high score regardless of whether they can actually do the work. Resume screening as a signal stopped working at exactly the moment hiring teams needed it most.

Third, the cost of one bad rejection compounds. A fresher rejected at the resume stage will not reapply — they tell three friends, post on Glassdoor, and never engage with that brand again. For consumer-facing companies hiring at campus scale, the brand cost of mass-rejection is now bigger than the recruiting cost.

What rejection-first screening looks like

The fix isn’t a better resume parser. It’s accepting that the resume is no longer the right artifact to screen on, and moving the first real filter one stage later — to a 15–20 min structured screening conversation with the candidate. A short AI voice screen at ₹59 per interview is cheaper than the human time burned by a botched resume filter, and it produces a signal a keyword match cannot: can the candidate hold a conversation, explain a real project, and respond to a follow-up question they didn’t prepare for.

We’ve run this pattern at scale across 1,099+ interviews in pilot campaigns, with single-evening drives that have screened 3,000 candidates in 2 hours. The lesson is consistent — a meaningful share of candidates rejected by the legacy resume screen are conversationally competent and pass communication and basic functional checks once you put a real voice screen in front of them. Those are the offers your old funnel was throwing away. (Why we screen communication first covers the broader argument.)

What this means for your funnel

If you’re running campus hiring at any meaningful scale in 2026, three numbers are worth measuring next quarter.

First, your resume-stage rejection rate. If you reject the majority of applications without a recruiter ever seeing the CV, the resume filter is doing too much of your decision-making, and almost certainly throwing out qualified people. You can validate this with a sampling exercise — pull a hundred resumes the ATS rejected, run them through a structured voice screen, and count how many would have made it past the next two human stages. That number is the size of your invisible leak.

Second, your second-look rate — what fraction of resume-rejected candidates actually clear a structured screen when you sample them. A healthy funnel has a low single-digit second-look rate. Anything substantially higher means your resume screen is filtering on the wrong signals.

Third, your candidate brand metric. If your reject-to-Glassdoor-complaint ratio is rising, the resume screen is silently torching your employer brand and you’ll feel it in next year’s application count.

The take

The next 18 months of campus hiring will be defined by which TA teams move the first real filter past the resume. The ones that do will hire from a larger qualified pool at lower cost-per-hire, with a better candidate experience to show for it. The ones that don’t will keep rejecting their best candidates at the gate and call it efficiency.

The resume isn’t a screen anymore. It’s a name tag. Stop using it to decide who gets a conversation, and start using a real conversation to decide who gets an offer.

See how a 15-minute voice screen replaces the resume bottleneck →