Overview

Lakshmi Machine Works (LMW), headquartered in Coimbatore, India, is one of the world's leading manufacturers of capital equipment for the textile and machining industries. LMW designs and manufactures machinery across the full yarn-spinning process as well as CNC machine tools and precision components through its Textile Machine Division (TMD) and Machine Tool Division (MTD). LMW holds an 84% market share in India across its textile machinery segment, a 60% market share in turnkey manufacturing setups, and operates 23 product lines across TMD and MTD. The company has a growing global footprint with customers and operations in China, Europe, and the Middle East. At the heart of LMW's business is a large and distributed workforce: field service engineers who maintain machines at customer sites, mill technicians who configure and operate complex spinning equipment, and machine operators who manage production quality on the floor. The performance and experience of this workforce directly determines LMW's ability to deliver on its promises to customers.

Supernova Award Category

Future of Work: Employee Experience

The Problem

LMW's workforce faced a structural problem: the knowledge required to operate, configure, and service its machines was deeply specialist, built over years of hands-on experience, and concentrated in a shrinking pool of senior engineers and master technicians. New operators took months to become proficient, service engineers resolved complex field issues by relying on senior engineer’s expertise, and mill technicians relied on institutional memory to configure machines for yarn quality -- with no systematic way to transfer that expertise to the next generation of workers. At the same time, operators on the shop floor were buried in nuisance alarms and reactive firefighting. LMW recognised that the only scalable answer was to embed intelligence into its machines and tools -- making expert knowledge accessible to every worker

The Solution

Altimetrik worked with LMW's Chairman and leadership team to assess the root causes of workforce knowledge dependency and design an AI programme that put intelligence directly into the hands of the people who needed it most. Six AI and ML solutions were deployed across LMW's TMD and MTD, each targeting a specific workforce pain point: Intelligent Service System (MTD and TMD): A GenAI-powered chatbot trained on LMW's full library of historical service cases, SOPs, and product manuals. Service engineers now query in natural language and receive precise, contextual resolution guidance in seconds -- replacing hours of manual search and reliance on senior colleagues. Yarn Quality Prediction (TMD): An ML web application that gives mill technicians AI-generated recommendations for optimal machine settings and raw material combinations, predicting key yarn quality metrics (CSP, RKM, U%, IPI) with approximately 90% accuracy. Operators who previously needed a master technician to make quality calls can now act on AI guidance independently. Alarm Severity Prediction (TMD): A real-time ML classification system that filters machine alarms at source, distinguishing critical faults from nuisance stops and sending STOP, WARN, or IGNORE commands directly to the controller. Operators are freed from responding to alarms that require no human action, reducing cognitive load and restoring focus to meaningful work. Thermal Error Compensation (MTD): A closed-loop ML system that predicts and compensates for thermal errors in CNC machines autonomously, with sub-second feedback. Machine operators no longer carry the burden of manual thermal correction across production shifts. Tool Wear Prediction (MTD): An ML model that classifies tool and insert condition in real time on the HMI. Operators can see machine health at a glance and act before failures occur, shifting from reactive intervention to informed decision-making. Sliver Quality Diagnostics (TMD): A diagnostic ML analysis that captures high-resolution sensor signals to detect roller timing mismatches during start-up and recommends corrective action. Technicians receive a precise diagnosis and fix path rather than needing to diagnose the issue from first principles.

The results

Before this programme, expertise was a bottleneck. New operators were exposed to machine complexity without adequate support. Service engineers resolved field issues through experience and memory. Mill technicians guarded quality configuration knowledge that was hard to document and harder to transfer. After deploying Altimetrik's AI solutions, expertise is embedded in the tools every worker uses. A service engineer with two years of experience now has access to the same resolution intelligence as a twenty-year veteran. A mill technician can configure machines for optimal yarn quality using AI guidance. An operator on the shop floor no longer spends the day responding to alarms that the system now handles. The Intelligent Service System produced measurable workforce outcomes: reduced Mean Time to Repair (MTTR), improved First-Time Fix Ratio, and a sustained increase in daily ticket closures. These numbers represent a direct improvement in the working experience of field service engineers, who now spend less time searching and more time solving. The alarm suppression system reduced nuisance machine stops by approximately 93%, directly cutting the volume of unproductive interruptions that operators faced each shift. This is one of the most concrete improvements to day-to-day operator experience achievable through AI. Looking ahead, LMW is building 23 AI products -- one for each of its product lines -- systematically extending these workforce benefits across every machine it manufactures.

Metrics

Workforce Productivity: Intelligent Service System: Reduced MTTR, improved First-Time Fix Ratio, increased daily ticket closures per service engineer Alarm Severity Prediction: 93% reduction in nuisance machine stops -- direct reduction in unproductive operator interruptions per shift Operator Empowerment: Yarn Quality Prediction: approximately 90% accuracy in predicting key quality metrics (CSP, RKM, U%, IPI) -- operators act on AI guidance without needing specialist escalation Thermal Error Compensation: 80% reduction in thermal positioning error with sub-second autonomous correction -- operators relieved of manual thermal management Tool Wear Prediction: 80% classification accuracy across material types -- operators shift from reactive fault response to proactive machine health monitoring Business Outcomes Enabled by Workforce Transformation: Approximately 30% reduction in material wastage through higher machine precision Approximately 50% reduction in equipment downtime through faster issue resolution Approximately 90% accurate mill configuration, enabling consistent output quality

The Technology

The technology programme was designed around a core principle: every AI system should make a specific worker more effective at a specific task. Each solution was built to integrate into the existing workflow of the person using it, with no requirement for workers to change tools or learn new interfaces beyond what was necessary. The Intelligent Service System was built on a Generative AI architecture using Retrieval-Augmented Generation (RAG). It was trained on LMW's complete repository of historical service cases, SOPs, and product manuals, and deployed as a conversational chatbot interface. Service engineers interact with it in natural language, as they would with a knowledgeable colleague, and receive grounded, source-backed recommendations. The closed-loop use cases -- Thermal Error Compensation, Tool Wear Prediction, and Alarm Severity Prediction -- used ML models trained on LMW's operational data and deployed on edge hardware integrated with PLC boards and FANUC controllers. These systems work autonomously in the background, reducing the operational burden on machine operators without requiring any additional action from them. The Yarn Quality Prediction system was deployed as a web application on AWS, with a chatbot interface designed for mill technicians. It was trained on LMW's historical production data and provides actionable configuration recommendations in plain language, making specialist knowledge accessible to workers at any experience level. The Sliver Quality diagnostic system used high-resolution encoder and LVDT sensor signals to detect mechanical mismatches during machine start-up. Rather than presenting raw data, the system delivers a specific corrective recommendation, reducing the diagnostic burden on technicians and enabling faster, more consistent resolution.

Disruptive Factor

The assumption in manufacturing that workforce quality is determined by years of hands-on experience and the slow accumulation of specialist knowledge -- a model that does not scale and creates a permanent dependency on a small number of expert individuals. This programme challenged that assumption directly by encoding decades of institutional knowledge into AI systems that any worker can access from day one, making expertise a property of the organisation rather than of specific people. The implementation required building the AI foundation entirely from scratch -- LMW had no prior ML infrastructure, fragmented data across legacy sensors and controllers, and no established processes for capturing and structuring service knowledge -- while simultaneously delivering near-term outcomes that workers could see and use. As LMW scales this to 23 AI products and extends into uptime ownership, control tower services, and AI-augmented roles, it is redefining what employees do in manufacturing industry.

Shining Moment

In 2026, LMW launched its AI-powered Textile Machine Division capabilities at ITMA Singapore, one of the most prestigious global exhibitions for textile machinery. Altimetrik was part of that launch as the AI and technology partner behind the transformation.