Custom HRMS for Indian Manufacturing: How We Manage 71,000+ Employees With Biometric Integration

The HR problem in Indian manufacturing is not the same as the HR problem in a software company.

A 500-person technology firm needs a clean leave management interface, a payroll engine that handles RSUs and variable components, and a good onboarding workflow. These are solved problems. Dozens of off-the-shelf tools handle them well.

A 5,000-person manufacturing unit in Karnataka needs to track 1,800 permanent employees, 2,400 contract workers from fourteen different contractor companies, biometric attendance tied to ESSL hardware at six gate points, Professional Tax deductions that vary by state (the facility has workers whose home state differs from the work state), PF compliance for contract workers where the contractor is nominally the employer, overtime calculations under the Factories Act (not a 1x multiplier — it varies by hours worked and applicable state rules), and a muster roll that satisfies both the factory’s own records and the contractor billing team.

These are not the same problem. And the off-the-shelf tool that solves the first problem will not solve the second.

Where Standard HRMS Platforms Break for Manufacturing

We have evaluated dozens of HR software platforms in the context of manufacturing enterprise requirements. The failure modes are consistent:

Contract workforce is an afterthought. Most platforms have a concept of “employee” — a permanent staff member with a CTC, a PAN, a bank account, and a joining date. Contract workers — who may number more than the permanent workforce at a manufacturing facility — do not fit this model. They are employed by a contractor, paid by the contractor, but their working hours are tracked by the facility. The billing relationship between the facility and the contractor company requires a separate data flow that standard platforms either do not handle or handle very poorly.

Biometric hardware integration is absent or fragile. Indian manufacturing facilities overwhelmingly use ESSL, Matrix COSEC, ZKTeco, or similar biometric hardware for attendance. Standard HRMS platforms offer theoretical integrations — an API that the hardware manufacturer may or may not have implemented, with a sync frequency that is typically batch rather than real-time. When attendance data from the hardware does not match the HRMS, which is the system of record? The answer in most implementations is “it depends, and it requires a manual reconciliation.”

Multi-state payroll compliance is hard-coded rather than configurable. Professional Tax rates, Labour Welfare Fund contributions, and some ESI applicability thresholds vary by state. An HRMS that handles Karnataka PT correctly may calculate it incorrectly for workers whose home state is Maharashtra. A platform built and maintained by a team in one state, without manufacturing clients distributed across multiple states, will have gaps.

Statutory report formats are incomplete. PF returns, ESI returns, PT returns, and factory compliance registers have prescribed formats that change periodically. A platform that generates a report that is 80% right still creates significant manual work for the compliance team to complete the remaining 20% before submission.

What We Built for Hindalco

The HRMS we built and maintain for Hindalco Industries manages 71,000+ employees across their manufacturing locations in India. The architecture decisions that made this work:

Unified data model for permanent and contract workforce. Rather than treating contract workers as a second-class record type, we built a data model where every worker — permanent, fixed-term, and contract — has a complete attendance record, a complete working hours record, and a complete compliance record. The billing relationship (facility-to-contractor) is a separate but linked data layer. This means the muster roll for statutory compliance and the contractor billing statement are both generated from the same underlying data, without any manual reconciliation step.

Real-time ESSL hardware integration. We built direct integration with ESSL devices using the ESSL SDK, pulling punch data in real time rather than on a batch schedule. Every biometric punch — in, out, or exception — is written to the HRMS within seconds of the hardware event. The HRMS is always current. Attendance exceptions (late arrival, early departure, missed punch) trigger WhatsApp alerts to the worker’s supervisor immediately, not the following morning.

Full statutory payroll engine. The payroll engine covers PF (employer and employee contributions, including cases where the employer contributes on restricted wages while the employee contributes on actual wages), ESIC (applicable or not based on salary threshold and state), Professional Tax (rate tables for all 28 states and 8 Union Territories, updated with each state budget), Gratuity (15/26 formula with qualifying service verification), Bonus (Payment of Bonus Act calculation with minimum and maximum constraints), Labour Welfare Fund (state-specific rates), and TDS (Form 16 generation with IT Act section 10 exemptions). This is not a simplified payroll engine — it handles the edge cases that manufacturing payroll teams deal with every month.

Contract workforce billing module. Every contractor company has a contractor portal where their site supervisor logs daily attendance for their workers. The HRMS validates this against the biometric data from the facility’s ESSL hardware. At month end, the contractor billing is automatically calculated — days worked, applicable rate per category of worker, deductions for statutory contributions managed by the facility — and a billing statement is generated. The contractor approves and the facility’s accounts team processes it. What was previously a 5-day manual exercise at month end is now a 4-hour verification and approval.

Five Things Every Manufacturing HRMS Must Handle in 2026

Based on our experience building and maintaining HRMS for large Indian manufacturing enterprises, these are the capabilities that distinguish a genuine manufacturing HRMS from a repurposed office HR tool:

1. Biometric hardware integration must be real-time and hardware-agnostic. Batch sync creates data gaps. When a worker clocks in at 8:47 AM, the attendance record should exist in the system by 8:48 AM, not at 6:00 PM when the nightly sync runs. Hardware-agnostic means the system should work with ESSL, Matrix COSEC, ZKTeco, and Honeywell devices — manufacturing facilities change hardware vendors and should not be locked into a single hardware ecosystem.

2. Contract workforce management must be native, not an add-on. The contractor workforce typically represents 40–70% of the total headcount at Indian manufacturing facilities. Any HRMS that treats contract worker management as an optional module or a workaround will fail in practice.

3. All-states statutory compliance must be maintained and updated. PT rates, LWF rates, and ESI thresholds change with state budgets. The compliance team should not be manually tracking budget notifications and updating rate tables. This is the platform vendor’s responsibility.

4. Offline capability is not optional. Gate biometric readers lose network connectivity. Mobile attendance apps are used in factory areas with poor signal. The system must queue locally and sync when connectivity is restored. An attendance system that goes offline when the network goes down is not a manufacturing-grade system.

5. Audit trail for factory compliance registers. The Factories Act and applicable state rules require specific registers — Form 12 (annual leave register), Form 15 (register of overtime), Form 25 (register of adult workers). These must be generatable on demand for factory inspector visits, with complete historical data traceable to the underlying attendance and payroll records.

ANKASTRA: The Manufacturing HRMS We Now Offer as a Product

The architecture developed through the Hindalco and other manufacturing HRMS engagements is the foundation of ANKASTRA — Ajinkya Technologies’ AI-powered Attendance and Workforce Management System.

ANKASTRA extends the manufacturing HRMS foundation with AI face recognition attendance (using MobileFaceNet embeddings, deployed on kiosk hardware for contactless gate attendance), visitor pre-registration with QR gate passes, and a multi-tenant SaaS architecture that makes it accessible to manufacturing enterprises of all sizes — from a 200-person precision engineering MSME to a 50,000-person integrated steel complex.

If you are a manufacturing HR manager or plant operations head evaluating attendance and payroll systems, we are happy to show you how ANKASTRA handles the specific workflows of your facility type.

Amey Kadle is Founder and CEO of Kadle Global Pvt Ltd. Explore ANKASTRA → | Book a demo →