High-volume hiring

You were handed a number and a date. The applications are already arriving faster than anyone can read them. a number and a date

500

seats to fill against one joining date

It is the Monday after the drive. The inbox holds more applications than your team can open all week, and the people worth ringing today have three other calls by Thursday.

Talk to us about your ramp
High-volume hiring on Recruitkar

What 500 seats actually looks like

Five hundred seats is not one req. It is a queue of applicants somebody has to open, judge and decide about, then reach before they go quiet. Working later does not move the number. Each mark below is one applicant, and the real pile runs off the bottom of this screen.

Each mark is one applicant, on one req. The solid ones are the share anyone actually gets to.

6

ways candidates arrive, one ranked de-duplicated pipeline out

10

dimensions, roughly, each one is scored on against your actual JD

0.5

credits to screen a resume already sitting in your folder

Three things that break at this volume

None of these break on a five-seat req. All three break on this one, and each has a point where it gives way.

by application 400

Reading stops being reading

The first eighty get read. The next eighty get skimmed. After that the shortlist is whatever was on top of the pile at 7pm. Nobody decides to stop reading, it just stops, and Monday's batch gets filled out of the leftovers.

by day three

The ones worth calling go cold

Someone worth a call on Monday has three other calls by Wednesday. At this volume the gap between shortlisted and contacted is hours, not days. You already know who to ring. That list is sitting behind 900 profiles nobody has opened, and every day lost here lands on a 60-day notice you are already counting backwards from.

by the third source

One person becomes three candidates

The extension pulls a profile off Naukri. The same person applies again through the drive link. Then they land a third time on the CSV from Saturday's walk-in desk. Three cards, three recruiters ringing the same number, two different verdicts, and the client sees the same person twice.

How a ramp actually runs

Four phases, each with a window it has to finish inside, because the joining date does not move for anybody.

day 0

Set the req up so the scoring means something

The job sits under a client and a department, so 500 seats across three profiles do not turn into one pile. Paste the JD you actually have. If it is two lines, Recruitkar tells you the score will be thin rather than handing you 500 confident-looking numbers built on nothing.

day 1-3

Fill the top

Everything you already have goes in at once: the resume folder, the extension on Naukri, Indeed and Shine, the walk-in CSV, LinkedIn, and the apply link you put on the drive poster. AI discovery across external profile databases covers what is still missing, at 1 credit a candidate, refunded when nobody is found. Each arrival is scored against that JD as it lands, and the same person arriving three ways becomes one card.

day 2-7

Reach them before they go cold

Bulk email with merge fields for the top slice, WhatsApp templates for the drive date and the reconfirm, AI voice calls for the ones who answer neither. Email is charged 1 credit per candidate ever, not per message, so the second and third follow-up cost nothing. Contact lookup charges 1 credit only when a number is actually found.

day 4-14

Interview at the volume, not below it

AI voice or video interviews in 11+ languages, recorded, with proctoring events, a scorecard and a hiring recommendation for each person. Your panel slots go only to people who already cleared that. Everyone else moves across the seven stages, so at 6pm on Friday you can say how many are offered, how many are hired, and which offers dropped this week.

Four phases, each with a window it has to finish inside, because the joining date does not move for anybody.

What holds when the volume triples

One person, one card, across every source

Six intake routes, one ranked pipeline. Nobody rings the same candidate twice because they came off Shine and the apply link. And if that candidate fits another of your reqs, putting them there is free and re-scores them against the new JD instead of carrying over an old verdict.

Every verdict shows its work

Matched skills and missing skills named one by one. Red flags. Interview questions suggested for that specific person. A recruiter who joined the drive yesterday reads the same reasons you would, which is the only way a bar survives being applied by eight people at once.

Eight recruiters, one budget

Per-member credit caps, so a new recruiter cannot burn the month on day two. Internal review can hold a member's candidate until a teammate approves it. When you share with the client, you pick which fields they see, and their approve, reject or note lands back on the candidate instead of in somebody's inbox.

The parts you are right to worry about

Half our volume arrives two or three times, extension plus drive link plus walk-in CSV. Where does that cost us twice?

Not in the work. They collapse into one ranked card, so your team rings once and the client sees one verdict. What you pay is the rate on each pull: resume screening 0.5, board import 0.5, LinkedIn import 1, sourcing 1 and refunded if the person is not found or comes back below the quality threshold. Email is 1 credit per candidate ever, not per message, and moving that same candidate onto another of your own reqs is free, with a fresh score against that JD.

Can it write 4,000 emails that do not read like 4,000 emails?

No. Know this before you buy. Email here is a blast tool: bulk send, merge fields, a deliverability pool, bounce monitoring. There is no AI writing a different email per candidate, and no domain warming bundled in. When one person needs a real conversation, that is what WhatsApp sessions and voice calls are for, at 2 credits per connected minute.

The client sent a two-line JD and wants 200 people by the 15th. What do the scores mean then?

Less than you want, and Recruitkar says so rather than hiding it. A thin JD, or a CSV row with only a name and an email, comes back with those dimensions empty instead of inflated. That is a real limit: vague JD in, vague verdicts out. The fix is ten minutes on the JD, not a better model, and you are told which parts are missing.

Most of our applicants will not interview comfortably in English, and we have no panel time for 500 of them.

AI interviews run in 11+ languages, voice or video, without booking a panel slot, each one recorded with proctoring events, a scorecard and a hiring recommendation. It is a recommendation. You still set the bar and you still decide who joins the batch. Billing is 4 credits a minute capped at 60, so one long interview cannot quietly run away with the budget.

You already have the number and the date.

Bring the req, the JD and whatever list exists today: the resume folder, the open Naukri tabs, Saturday's walk-in CSV. We will set the first drive up with you.

Talk to us about your rampSee how credits work