Case · Build (self-applied) · 11 weeks (Build-shaped, internal)
Internal — Talpro Universe · IT staffing recruiter desktop
Build (self-applied) · 11 weeks (Build-shaped, internal)
Talpro India
Outcomes · measured
Shortlists per recruiter / week
+50% · sustained 9 weeks post-rollout
24h SLA hit rate
across 3 demanding clients
Inference cost
12-recruiter load · <2% of wage bill
Narrative
01 · Problem
Problem
Talpro India's 12-recruiter desk in Bengaluru was hitting its productivity ceiling. Average recruiter handled 90 active roles and shipped 18 shortlists a week — the headcount-vs-output ratio that had carried the firm for a decade. By Q1 2026, two pressures collided: the firm needed to grow placements 40% YoY without adding recruiters, and three large clients were demanding 24-hour shortlist turnaround SLAs that the existing manual process couldn't hit. Three off-the-shelf recruiter-AI tools had been trialled (Manatal, Recruitee AI add-on, BizMerlinHR) and dropped within 60 days each. The pattern: each tool indexed candidates well in isolation but couldn't see the recruiter's side conversations, the existing HubSpot deal pipeline, or the WhatsApp-based candidate negotiation log that lived outside the ATS. Recruiters routed around the tools within two weeks; tools became expense, not leverage.
02 · Approach
Approach
HireIQ Pro built in 11 weeks as a recruiter-shaped desktop, not a candidate-shaped one. Week 1: shadow three recruiters end-to-end; document the 41 micro-tasks that fill a recruiter day. Weeks 2–3: define the seven that AI could legitimately compress (intake parsing, shortlist drafting, JD-to-CV match scoring, follow-up cadence, status-update drafting, candidate-objection-bank lookup, week-close report). Weeks 4–8: ship those seven as a single Next.js desktop with HubSpot, WhatsApp Business API, and Snov.io enrichment wired in. TALPRO-IQ scored shortlist-rank with a distilled 7B model first-pass and Claude Sonnet 4.5 for top-20 refinement. Weeks 9–10: rollout to 4 of 12 recruiters with a 2-week A/B vs the legacy process; daily recruiter-feedback huddle. Week 11: full rollout once the A/B showed shortlists-per-week up 38% with recruiter-agreement on AI rankings stable above 0.78.
03 · Outcome
Outcome
Shortlists-per-recruiter-per-week rose from 18 to 27 (+50%) on identical role mix, sustained for 9 weeks post-rollout. The 24-hour SLA hit 92% on-time across the three demanding clients (was failing at 41%). Recruiter-AI agreement on shortlists settled at 0.81. Total HireIQ Pro inference cost ran ₹14,200/month at full 12-recruiter load — under 2% of the recruiter wage bill it amplifies. The platform now serves as the playbook the Agent Army Sales-Dev squad uses for client deployments: same pattern (desktop-shaped, not tool-shaped), same eval harness, same rollout cadence.
04 · In their words
“We tried three AI recruiter tools that were built around the candidate. HireIQ Pro is built around the recruiter. That’s the difference between expense and leverage.”
05 — Who led this engagement
Bhaskar Anand. Every first call.
Founder & CEO, CompetitorX. Pune, India. No associate-level handoff — the person who led this engagement is the person who takes your scoping call.