For the agency recruiter
You get paid when someone joins. Everything before that is unbilled.
It’s the 24th. Two offers are out, one candidate has stopped replying, and you’re still sourcing for a req that opened on Monday. Recruitkar takes the part of that week that was never going to earn you anything.
Your week, as it actually goes
Not the version in the tracker. The version where four hours disappear and you cannot point at what took them.
Monday, 9am
A new req lands. You already know two other agencies have it, and the one who gets a credible shortlist in first is the one who gets paid. So you start the same search you have run forty times before, from the beginning, again.
Wednesday, 4pm
Twenty CVs open, renamed, half-read. Between them: two client calls, a candidate who wants to reopen the CTC conversation, and a manager on another req asking where things stand. You answer from memory, because that is where the pipeline lives.
Friday, 7pm
You send five profiles. Then the silence — the client will come back on Wednesday, and by Wednesday two of your five have taken something else. Not because they were not interested. Because somebody moved faster.
What you are actually judged on
Not activity. Three numbers, and one of them is your month.
1st
credible shortlist in, on a shared req
On contingency, second place pays nothing. Speed is not a nice-to-have here, it is the entire fee.
1 in 5
profiles that survive the client’s read
Send five, get one interview, and the hiring manager starts opening your next mail with less patience than the last.
24th
of the month, when the maths gets real
Offers out, joinings pending, and a target that does not care which of them drops in the last week.
What leaves your desk
The work that was never going to earn you a rupee.
Downloading and renaming CVs
Tick the applicants on your Naukri, Indeed or Shine tab and the browser extension brings them across at half a credit each, already de-duplicated against everyone else on the job. The folder of resumes from the weekend drive goes up in one go.
Reading forty profiles to find five
Every profile is scored against the client’s actual JD, with a verdict from strong fit to not fit, matched and missing skills named one by one, red flags, and the questions worth asking that particular person. You work the top of the list instead of the list.
Typing the same message sixty times
Bulk email with merge fields, WhatsApp templates, or an AI voice call, all run off the same ranked list. Email is one credit per candidate ever, not per message, so every follow-up after the first one is free.
Asking the same six questions until you are numb
Current CTC, expected, notice period, buyout, offers in hand, ready to relocate. An AI voice or video interview asks them in the candidate’s own language, records it, and leaves a scorecard and a recommendation on the card. It does not get tired at the fortieth call.
The part that stays yours
Straight answers
I get paid on placements, not on software. What does this cost me before I have billed anything?
Credits, spent on work that actually happens. Half a credit to import or screen a resume. One to source a candidate, refunded if they cannot be found and refunded again if the match comes back below the quality threshold. One per candidate you ever email, however many times you follow up after that. There is no per-seat licence to commit to for the year before you have placed anybody.
My candidates come off Naukri and my own old database. I am not going to start sourcing from somewhere I have never heard of.
Then don’t. Install the browser extension and pull in the applicants already sitting on your Naukri, Indeed or Shine req, or upload the folder you have had since March. They get scored exactly like anything else. AI discovery is there if you want it, not a gate you have to pass through to use the rest.
I have seen matching tools before. They give everybody eighty-five percent and I still end up reading all of them.
The honest limit first: the output is only as good as the JD you paste. Give it the two-liner the client sent on WhatsApp and it will not invent a confident number — those dimensions come back empty instead of flattering. You will see fewer scores, not more. Paste the real brief and the ranking is worth something, and you can see what it read either way.
Two of us work the same account and keep pulling the same person from different places.
Those arrivals merge into one card with one score and one history, because the pipeline de-duplicates across every source. He does not get called twice by two of your people, and he does not get the same blast twice — which is the version of this that actually costs you the account.
There is a req on your desk right now
Bring that one. The real brief, and the applicants already sitting on it. Look at the shortlist before you move anything else across.