Case Study · 04 / 10 Tech Recruitment Save as PDF ↓
CV screening + scheduling · 8-week build

The hiring funnel that screens itself.

How a Bangalore recruitment shop cut time-to-shortlist from 4 days to 6 hours — and doubled placements per recruiter.

ClientLattice Talent LocationBangalore, India PublishedMay 2026
Industry
Tech recruitment · Series A–C
Scale
80–100 open mandates · 3,200 CVs/wk
Team
62 employees · ₹18 Cr ARR
Engagement
8 weeks end-to-end
Outcomes · 90 days post-launch
4.2d → 6h
Time to first shortlist, from 4.2 days to 6 hours
2.7×
CVs processed per recruiter per month — 270 → 740
−44%
Cost per placement, ₹42,000 → ₹23,500
§ 01 · The company

Series-A-to-C placements at India + Singapore speed.

The company

Lattice Talent is a Bangalore-based tech recruitment agency placing senior engineers and PMs at Series A–C startups across India and Singapore. 62 employees, ~₹18 Cr ARR, founded in 2019. They run 80–100 open mandates at any time, mostly for SaaS, fintech, and AI companies that need senior IC hires fast.

Aarti Menon, COO and co-founder, was losing sleep: her best recruiters were spending more time reading CVs than talking to candidates, and three clients had escalated about slow shortlists in the same week.

§ 02 · The problem

3,200 CVs a week. 38% of recruiter time on screening admin.

Inbound load

Lattice was pulling in roughly 3,200 inbound applications a week across LinkedIn, Naukri, referrals, and their own talent pool. 12 recruiters did the first-pass screening manually — averaging 4 minutes per CV. Time from application to first shortlist sat at 4.2 days. Internal audits showed 38% of recruiter time went to screening admin instead of candidate calls or client work.

Breaking pointOctober

Lattice lost a Staff Engineer placement at a Series B fintech because the candidate accepted a competing offer on day three — before Lattice had even shortlisted him. A follow-up audit found 22% of A-tier candidates were going cold to other agencies inside the first 72 hours.

22% of A-tier going cold < 72h

Two senior recruiters had quit citing burnout in the previous quarter. Aarti was about to onboard three more bodies just to keep up with inbound.

§ 03 · The solution

A screening + scheduling layer on top of Recruiterflow.

Architecture

Vynara built a screening and scheduling layer on top of Lattice's existing ATS (Recruiterflow). Inbound CVs get parsed, embedded, and scored against a per-mandate rubric that senior recruiters helped design in week one — so the model scores the way the firm's best recruiter would.

Each candidate gets a 0–100 score plus a 3-line rationale: why they fit, where they fall short, and a confidence flag. Top 8% per role get auto-routed into a WhatsApp flow that confirms interest, captures a 90-second voice intro, and books a slot on the assigned recruiter's calendar. Humans only enter the loop when a candidate is already pre-qualified and warm.

Auto-route to WhatsApp only if
score  ≥  top 8% AND confidence flag = high
§ 04 · How it works

From 4 minutes per CV to 9 seconds.

Before · Manual screen 4 min / CV · 4.2d
  1. Application lands in ATS
  2. Recruiter opens CV, re-reads JD
  3. Scores manually, writes notes
  4. Emails candidate for availability
  5. 2–3 day email back-and-forth
  6. Schedules call, finally talks to candidate
After · Automated screen 9 sec / CV · 6h
  1. AI parses + scores against rubric9 sec / CV
  2. Top 8% auto-routed to WhatsApp
  3. Candidate confirms + records 90-sec voice intro
  4. AI books slot on recruiter's calendar (Cal.com)
  5. Recruiter walks into a warm call with a 3-line brief
§ 05 · The numbers

Six metrics. Same engagement.

Time to first shortlist−94%
Before4.2 days
After6 hours
Recruiter time on screening−76%
Before38%
After9%
CVs / recruiter / month+2.7×
Before270
After740
Placements / recruiter / quarter+1.7×
Before11
After19
Offer-acceptance rate+12 pts
Before61%
After73%
Cost per placement−44%
Before₹42,000
After₹23,500
§ 06 · The timeline

Eight weeks. Audit dashboard included.

Phase
W1
W2
W3
W4
W5
W6
W7
W8
Weeks 1–2Discovery · reverse-engineered rubric · ATS schema
DISCOVERY
DISCOVERY
Weeks 3–4CV parser · embedding pipeline · rubric scoring · backtest 4,000 CVs
BUILD
BUILD
Week 5Recruiterflow API · two-way sync
INTEGRATE
Week 6WhatsApp flow · Cal.com · voice intro capture
FLOW
Week 7Shadow mode · 87% human-AI agreement
SHADOW
Week 8Cutover · audit dashboard for override
LIVE
§ 07 · What they said

Same throughput. Fewer people. Better margins.

"

I stopped hiring the three recruiters I was about to put offers out to. Instead I gave my existing team a raise. Same throughput, fewer people, better margins.

AM
Aarti Menon
COO & Co-founder · Lattice Talent
"

I used to dread Monday mornings staring at 200 new CVs. Now I walk in and my calendar already has six warm calls booked. I am actually recruiting again.

RK
Rohan K.
Senior Recruiter · Lattice Talent
§ 08 · The takeaway

Delete the worst part of the job.

The win wasn't replacing recruiters — it was deleting the worst part of their job so they could do the part that actually closes placements.

Any agency or in-house TA team running 40+ open roles at a time has the same shape of problem and the same shape of fix. If your recruiters are reading CVs instead of talking to humans, that's the symptom worth solving first.

One-liner We helped a Bangalore recruitment agency cut time-to-shortlist from 4 days to 6 hours and double placements per recruiter.
Show us your CV pipeline →