The factory floor learns to think: AI jobs in Indian manufacturing

NITI Aayog wants manufacturing to reach more than 25% of India's GDP and create over 100 million jobs by 2035, and it names AI and machine learning, digital twins and robotics as high-impact enablers (PIB, 2025). Without frontier technology, it warns, India could lose $270 billion by 2035.

That makes AI a factory-floor skill, not only an IT one. This post covers what is happening, which roles gain, the skills to build, and what still needs people.

What is happening in Indian manufacturing

Adoption is real but early. In Deloitte's 2025 smart manufacturing survey, 29% of manufacturers use AI or machine learning at facility or network level, and 24% use generative AI (Deloitte, 2025). The same survey found 48% struggle to fill production and operations roles. The shortage is people, not machines.

One Indian plant shows what is possible. DCM Shriram Chemicals says its Jhagadia site in Gujarat runs 45 digital and analytics solutions, including AI process control and a generative-AI maintenance manager (The Wire, 2026). The company claims EBITDA rose 11 percentage points and power costs fell 32%. These are the company's own figures, but they show where the value sits: process, energy and maintenance.

Supply chains are moving too. Walmart says its Self-Healing Inventory system has saved more than $55 million (Walmart, 2025).

Entry-level work is where the change will be felt first. In a Gartner survey of 509 supply-chain leaders, 55% expect agentic AI to reduce entry-level hiring, and 86% say it requires new ways to develop talent pipelines (Robotics 24/7, 2026). Globally, data entry clerks are among the fastest-declining jobs (WEF, 2025).

Which roles gain, and how

The examples below are illustrations of a working day, not survey findings.

Maintenance technicians. A pump on line 3 starts vibrating at 2 a.m. The technician asks a maintenance assistant, trained on the plant's own manuals and repair history, what caused the same symptom last time. It points to a bearing and the part number. He checks, confirms, and fixes it. The assistant saved the hunt through a binder. The diagnosis is still his.

Process and quality engineers. A quality engineer uploads a week of defect reports and asks for patterns by shift, machine and raw-material batch. The assistant groups them. She notices that one supplier's batch lines up with the spike, and walks to the line to check.

Shift supervisors. Instead of writing a handover by hand, a supervisor dictates what happened on the shift and gets a clean summary for the next team, with open issues listed first.

Energy and utilities managers. Power is a large cost in process plants. A manager who can read an AI model's recommendation for load scheduling, and challenge it when it ignores a real constraint, becomes central to the plant's margins.

Supply-chain planners and procurement staff. A planner gets a draft reorder plan each morning. His job moves from typing numbers to questioning them: why is the model ordering less before a festival month?

Frontline operators. An operator reads a work instruction in her own language, generated from the English standard procedure, and flags where it doesn't match what the machine actually does.

The skills to build

• Asking good questions of plant data. You don't need to code. You need to ask a clear question and know when an answer doesn't match what you see on the floor.
• Checking AI advice against physical reality. A model can suggest a setting the machine cannot safely run. Microsoft's 2026 survey found workers rank quality control of AI output (50%) and critical thinking (46%) as the top human skills (Microsoft Work Trend Index, 2026).
• Documenting what you know. An assistant is only as good as the manuals and repair logs behind it. Senior technicians who write down what they know make the whole plant smarter.
• Teaching juniors on the job. If agentic AI reduces entry-level roles, the juniors who remain need faster, better coaching. In a study of 5,172 support agents, AI helped less-experienced workers most (Brynjolfsson, Li, Raymond, 2025).

The need is large. EY India estimates AI could transform 38 million Indian jobs by 2030 (EY India, 2025), and the WEF says 59 of every 100 workers will need training by 2030.

The risks, and what still needs humans

AI is uneven in ways that are hard to predict. In a Harvard and BCG study, consultants were right 84% of the time on a task designed for AI to fail, but only 60–70% of the time with AI (HBS AI Institute). On a factory floor, a confident wrong answer can damage equipment or hurt someone.

Losing the ladder is the quieter risk. If fewer freshers are hired, fewer people learn the plant from the ground up. In a randomised trial on learning a coding skill, people who used AI scored 50% on a follow-up quiz against 67% for those who didn't, though those who asked the AI for explanations kept learning (Anthropic, 2026). How you use the tool decides whether you grow.

What to do next

1. Write down the three questions you answer most often on a shift. Those are your first AI use cases.
2. Collect your plant's manuals, SOPs and repair notes in one place. Clean documents come before any useful assistant.
3. Try an approved assistant on one task, such as shift handovers, for two weeks, and check every output.
4. When you use AI to solve a problem, ask it to explain why, not only what. You will learn faster.
5. Pair with a junior colleague and teach one AI habit you have found useful.

If you want to build these skills with others, AI Fluency is taught live in small cohorts, with a project you keep each month.

Keep reading