AI and the Nine Wastes of Lean: Where the Information Gap Was Always the Real Problem

By Palaniappan SN · Co-Founder, StratAI · MBA, IIM Bangalore · BE (Mechanical), PSG Tech

Most manufacturing companies have tried lean. Many have tried 5S, Kaizen events, value stream mapping. Some have had consultants. The principles are sound — identify waste, eliminate it at source, repeat. The results, in most mid-market plants, have been partial. Lean sticks on the floor for a while and then the old ways return. The tools were right. The information infrastructure was not there to sustain them.

AI and lean are not two separate programmes. AI does not replace lean — it closes the information gap that lean always required but could never fill manually. For each of the Nine Wastes, lean identifies where the problem is. AI provides the data, the visibility, and the feedback speed to make the elimination stick.

Direct answer: AI helps in lean manufacturing by closing the information gap that prevents waste elimination from being sustained. Lean principles correctly identify the Nine Wastes — Defects, Waiting, Overproduction, Non-Utilised Talent, Transportation, Inventory, Motion, Overprocessing, and Extra Processing (information waste). For each waste, the lean solution requires data that most mid-market manufacturing plants do not have in real time: defect patterns connected to root causes, batch waiting times across stations, demand signals connected to inventory positions, and process data connected to machine and product specifications. AI makes this data available, makes the feedback loop faster, and enables human experts to act on insights that were previously invisible. AI is not the solution — it removes the data and knowledge barrier to the solution. The fundamentals remain: identify the waste, find the root cause, address it at source, repeat.

The Nine Wastes — Where AI Closes the Gap

Each waste block below shows: what lean says the problem is, where the information gap prevents elimination, and what AI specifically enables. Each block also shows implementation priority — because the wastes are not all equally addressable from day one.

Waste 01 · Defects

The lean view: Rejected parts, rework, scrap. Lean response: poka-yoke, standardised work, SPC at source.

Where AI closes the gap: Defects are a feedback speed problem before they are a technical problem. The root cause exists — in the machine, the product, the process, or the operator — but by the time the defect data arrives, the next batch has already run. AI closes the feedback loop: mobile image-based QC capture digitises defect data at the moment of inspection, not days later. AI trains on machine-specific, product-specific, process-specific, and operator-specific defect patterns and builds a knowledge base so that root causes are not discovered repeatedly — they are accumulated and acted on. Take aluminium die casting as an example. Material cost is the largest component of production cost. Defects are expensive — not just in scrap but in the machine time, energy, and labour already consumed on a rejected part. The defect data exists. It sits in QC sheets, shift reports, and inspection records — siloed, unconnected, and rarely acted on systematically. In one aluminium die casting operation we observed, the defect rate was running at 4%. Reducing it to 2.5% was achievable — not through a new machine or a new process, but through iterative action on the defect data that was already being generated. AI makes that iteration faster, more consistent, and cumulative across every subsequent run. The lean principle (identify, root cause, fix at source, repeat) remains. AI makes the cycle shorter and the knowledge permanent.

Implementation priority: Phase 1 — Start here

Waste 02 · Waiting

The lean view: Idle time — operators waiting for materials, machines waiting for jobs, batches sitting between stations. Lean response: flow, pull systems, takt time.

Where AI closes the gap: Waiting waste is invisible because it is never measured. Nobody records when a batch entered a station, how long it waited, or where the inventory built up. Without that data, bottleneck identification is guesswork. AI enables batch-level timestamping through photo-based capture — when a batch enters which machine, what the output was, who produced it, what the QC result was, all recorded in real time. With this data, a human expert can identify which batches waited the longest, which station is the inventory chokepoint, and which bottleneck — if addressed — would produce the greatest reduction in waiting time. The conventional solutions (better planning, quicker setups, bottleneck removal) remain correct. AI provides the visibility these solutions need to be targeted correctly.

Field observation: Waiting waste is currently invisible in most mid-market plants — not because it does not exist, but because nobody has the data to see it.

Implementation priority: Phase 1 — Start here

Waste 03 · Overproduction

The lean view: Making more than needed, before it is needed. Lean response: pull systems, kanban, takt time alignment.

Where AI closes the gap: For made-to-order manufacturing, overproduction is not the primary problem — pull is inherently built in. For made-to-stock, overproduction is driven by a demand forecasting failure: the sales data exists but nobody has calculated its distribution, standard deviation, or run a continuous replenishment model with proper service levels and optimal order quantities. The data sits unused. AI surfaces these patterns, builds demand visibility, and enables human experts to co-think with AI and iteratively improve forecasting accuracy over time. The improvement compounds — each cycle produces better data, which produces better forecasts, which produces less overproduction.

Implementation priority: Phase 2 — Build on data foundation

Waste 04 · Non-Utilised Talent

The lean view: Underuse of people’s skills, knowledge, and creativity. Lean response: employee involvement, suggestion systems, cross-training.

Where AI closes the gap: This is the most human of the nine wastes — and the hardest to address without changing the management environment. Young teams in mid-market manufacturing want challenge, freedom to fail, and problems worth solving. What they typically get is repetitive work with no visibility into whether it matters. AI unlocks two things that change this: knowledge (what was previously locked in consultants, senior operators, and undigitised experience becomes accessible to the entire team) and visibility and feedback (people can now see the impact of their decisions in real time). Small, solution-oriented teams guided by knowledge and empowered by AI can solve core problems over time. But the management environment must exist first — conventional, slow, hierarchical structures change nothing regardless of what tools are available. AI is the enabler. Leadership is the prerequisite.

Field observation: In a recent meeting with a mid-market company’s inventory planning team: all young, capable people. Wanting challenge. But no visibility, no knowledge infrastructure, no problem to own. The waste was not in the people. It was in the environment.

Implementation priority: Phase 1 — Enable people

Waste 05 · Transportation

The lean view: Unnecessary movement of materials between locations. Lean response: facility layout optimisation, co-location, value stream mapping.

Where AI closes the gap: Transportation waste is a layout problem — it is not dynamic and continuous like defect or waiting waste. The AI contribution is in the analysis phase: when real production data (batch movement, bottleneck stations, inventory build-up points) has been captured over time, AI can surface patterns that inform layout decisions during plant expansions or redesigns. Co-thinking with AI on better layout planning using accumulated production data is a high-value application — but it requires the foundational data layer to exist first. New plant or expansion planning is where this pays off most directly. We have not deployed a transportation waste AI system specifically, but the production data from Phase 1 — batch movement, bottleneck stations, inventory build-up points — makes layout analysis straightforward to conduct during expansion planning.

Implementation priority: Phase 3 — Later stage use case

Waste 06 · Inventory

The lean view: Excess raw material, WIP, or finished goods. Lean response: JIT, kanban, pull systems.

Where AI closes the gap: Inventory waste and overproduction waste are two sides of the same information gap — both driven by demand forecasting quality and sourcing discipline. The current mindset in most mid-market plants: delivery is the metric, inventory is ignored. As long as orders ship, excess stock is invisible as a problem. AI changes this by creating continuous feedback between demand signals, inventory position, and sourcing decisions — and by enabling simulations of how demand can be served at optimal inventory levels. Preventive alerts when inventory drops below reorder point, or builds above buffer level, change inventory from a static management problem to a dynamic one. Built ground up from iterations — not a day-one system, a compound improvement over time.

Implementation priority: Phase 2 — Build on data foundation

Waste 07 · Motion

The lean view: Unnecessary movement of people — reaching, walking, searching. Lean response: 5S, ergonomic workstation design, shadow boards.

Where AI closes the gap: Motion waste is a physical design problem. Like transportation, the AI contribution is in the analysis phase — image capture of workstation layouts, movement pattern analysis, co-thinking with AI to design better ergonomics and tool placement. Once the improved workstation design is validated on one line, replicate it across the plant. This is a later-stage use case — valuable once the foundational data capture and production intelligence from Phase 1 exists. Like transportation, this is a design problem that benefits from accumulated production data rather than a real-time AI system.

Implementation priority: Phase 3 — Later stage use case

Waste 08 · Overprocessing

The lean view: Doing more than the customer requires — extra steps, redundant checks, unnecessary precision. Lean response: understand true customer requirements, eliminate non-value-adding steps.

Where AI closes the gap: Overprocessing is a specification and standards problem. Comparing the spec sheet and ballooning diagram against the actual process — and co-thinking with AI to identify where effort exceeds what the customer needs — is a legitimate and valuable application. But it requires the foundational data layer to already exist, and it is not a pressing problem for most mid-market manufacturers where defect, waiting, and inventory wastes are consuming far more cost. Important eventually. Not where to start.

Implementation priority: Phase 3 — Later stage use case

Waste 09 · Extra Processing — Information Waste

The lean view: Unnecessary processing of information — duplicate data entry, manual report generation, information handled multiple times. Often called the ninth waste.

Where AI closes the gap: This is the most directly and immediately addressable of all nine wastes with AI — and the most underestimated. The conventional assumption is that ERP implementation reduces manual data work. The field reality is frequently the opposite. AI reverses this: image-based capture at source eliminates manual entry, information reaches the right decision-maker at the right moment in the right format, and the data entry headcount shrinks rather than grows. This is a Phase 1 use case — and in most mid-market plants, it is the use case that unlocks every other AI application by creating the data infrastructure the other wastes require.

Field observation: One company went from 7-8 data entry workers to more than 25 after ERP implementation — because the system created more data entry requirements than it eliminated. AI-based capture reverses this entirely.

Implementation priority: Phase 1 — Start here

AI integration with lean tools reduced non-value-added activity from 75.3% to 59.5% and cut lead time by 12 days. Lean identifies where the waste is. AI makes the elimination systematic.

The combination works because lean and AI address different parts of the same problem. Lean provides the framework for identifying and eliminating waste. AI provides the data infrastructure that makes identification precise and elimination sustainable. Neither is sufficient alone — lean without data produces improvements that revert; AI without lean principles produces data without a framework for action.

Source: E3S Web of Conferences — Application of Lean Manufacturing to Minimise Waste, 2024

The Implementation Sequence — Not All Nine at Once

The nine wastes are not equally addressable from day one. The right sequence is:

• Defects — Phase 1, Start here — Mobile QC capture, root cause knowledge base, real-time feedback loop
• Waiting — Phase 1, Start here — Batch timestamping, bottleneck visibility, production planning intelligence
• Extra Processing (Information) — Phase 1, Start here — Image-based data capture, eliminate manual ERP entry, right info at right time
• Non-Utilised Talent — Phase 1, Enable people — AI as knowledge and visibility layer; small solution-oriented teams empowered
• Overproduction — Phase 2, Build on data — Demand forecasting, continuous replenishment model, inventory simulation
• Inventory — Phase 2, Build on data — Linked to demand forecasting; inventory-delivery feedback loop; preventive alerts
• Transportation — Phase 3, Later stage — Production data informs layout decisions during plant expansion
• Motion — Phase 3, Later stage — Workstation design optimisation using production and image data
• Overprocessing — Phase 3, Later stage — Spec vs process analysis; requires foundational data layer first

Start with the wastes that are data-starved and high-cost: Defects, Waiting, and Information Processing. These three create the data foundation that makes every subsequent waste reduction possible. Non-Utilised Talent is addressed simultaneously — not as a technology problem but as a leadership and knowledge access problem. Overproduction and Inventory follow once the demand and production data exists. Transportation, Motion, and Overprocessing are the final layer — valuable, but dependent on everything that came before.

Enterprises that redesigned workflows with AI reported a 63% boost in productivity — the highest impact came from AI applied to process improvement, not technology replacement.

The manufacturing companies that see the highest AI return are not the ones that deployed the most AI tools. They are the ones that deployed AI in service of an existing improvement discipline — lean, in this case. AI applied without a lean mindset produces local optimisations. AI applied with a lean mindset produces systematic waste elimination. The two are designed to work together.

Source: ServiceNow / Pearson, India AI Skills Report, 2025

AI in manufacturing in India is forecast to reach INR 12.59 billion by 2028 — a 58.96% CAGR. The fastest-growing adoption is among companies moving AI from pilots to core business functions.

The mid-market manufacturers that integrate AI with their existing lean programmes now are positioning for compounding returns. Each waste eliminated creates the data infrastructure for the next elimination. The manufacturers that wait will find the category more contested and the training data advantage already held by early movers.

Source: ResearchAndMarkets, AI in Manufacturing in India, 2024

What This Means for a Manufacturing CEO

If your lean programme has stalled — if 5S was implemented and then gradually abandoned, if the Kaizen events produced enthusiasm but not sustained change — the problem is almost certainly not the lean principles. The principles are correct. The problem is the information gap that lean alone cannot close.

AI does not ask you to abandon lean. It asks you to give lean the data infrastructure it always needed. Real-time defect feedback instead of monthly rejection reports. Batch-level waiting time visibility instead of guesswork about bottlenecks. Demand-linked inventory signals instead of production schedules built on last quarter’s sales. Information waste eliminated at source instead of managed by 25 people entering the same data into the same ERP in different ways.

The companies that will sustain lean improvement in the next decade are not the ones with the most sophisticated lean tools. They are the ones that connect lean discipline to AI-powered information flow. The waste was always there. The visibility was not.

The audit tells you which waste to address first — and which AI use case closes the information gap fastest.

Tell us which of the Nine Wastes is costing your plant the most right now.

→ Book your free half-day audit — no commitment, no strings

We map the specific waste with the highest cost in your operation and identify the first AI use case that closes the information gap. We confirm your audit date within one business day.

Frequently Asked Questions

How can AI help in lean manufacturing?

AI helps in lean manufacturing by closing the information gap that prevents waste elimination from being sustained. Lean principles correctly identify the Nine Wastes and the solutions for each. What lean alone cannot do is provide the real-time data infrastructure those solutions require — defect patterns linked to root causes, batch waiting times across stations, demand signals linked to inventory positions. AI makes this data available, accelerates the feedback loop, and enables human experts to act on insights that were previously invisible. AI is not the solution — it removes the data and knowledge barrier to the solution. The lean fundamentals remain: identify the waste, find the root cause, fix it at source, repeat.

What are the Nine Wastes of Lean Manufacturing?

The Nine Wastes of lean manufacturing are: Defects (rejected parts, rework, scrap), Waiting (idle time between process steps), Overproduction (making more than needed), Non-Utilised Talent (underuse of people’s skills and creativity), Transportation (unnecessary material movement), Inventory (excess raw material, WIP, or finished goods), Motion (unnecessary movement of people), Overprocessing (doing more than the customer requires), and Extra Processing — Information Waste (duplicate data entry, manual report generation, information handled multiple times). The ninth waste — information processing — is the most directly addressable with AI and is often the one that unlocks every other waste reduction.

Why does lean manufacturing fail to stick in mid-market plants?

Lean fails to sustain not because the principles are wrong — they are correct — but because the information infrastructure required to sustain it does not exist. 5S improves the floor but the Waiting waste continues because the production schedule does not reflect this morning’s priority change. Defect reduction initiatives produce results but revert because the QC data from yesterday’s shift is not connected to today’s operator’s decisions. Lean requires feedback speed that manual systems cannot maintain. AI provides that feedback speed — which is why AI and lean are complementary, not competing.

Which lean waste should we address with AI first?

Start with three wastes simultaneously: Defects, Waiting, and Information Processing (the ninth waste). These three are the most data-starved and the most directly addressable with AI from day one. Defect waste requires real-time QC capture and root cause mapping. Waiting waste requires batch-level timestamping and bottleneck visibility. Information processing waste requires AI-based data capture to eliminate manual ERP entry. These three create the data foundation that makes every subsequent waste reduction — Overproduction, Inventory, Transportation, Motion, Overprocessing — possible to target precisely.

Can AI replace lean manufacturing consultants and tools?

No. AI and lean consultants serve different functions. A lean consultant brings the framework for identifying waste, the change management discipline to sustain improvement, and the experience to know which waste is causing the most cost in this specific plant. AI provides the data infrastructure, the feedback speed, and the knowledge accumulation that makes lean principles work continuously rather than in periodic events. The combination — lean discipline + AI-powered information flow — produces better results than either alone. AI without lean principles produces local optimisations without a systematic improvement framework. Lean without AI produces correct diagnoses without the data infrastructure to sustain the cure.

About StratAI

StratAI helps manufacturing firms in India build AI Advantage Systems. 10+ live deployments across textile, jewellery, furnishings, commodity processing, and component manufacturing. Official Registered Claude Partner and Anthropic Partner.

stratai.io/contact · palani@stratai.io · +91 99402 25924

“Lean identified where the waste was. AI closes the information gap that was always preventing you from eliminating it.”

— Palaniappan SN, Co-Founder, StratAI