# AI plans spare parts stock for after-sales service: less inventory, fewer backorders (Mitsubishi Electric, HiKOKI, Metso)

> AI service parts planning: Mitsubishi Electric Europe cut spare parts stock 30%, service 87% to 97%; HiKOKI cut backorders from 4 months to 4 weeks.

- Canonical URL: https://aiagentsformanufacturing.com/use-cases/ai-service-parts-inventory-planning-aftermarket-mitsubishi-electric-metso/
- Department: Logistics / Supply Chain (https://aiagentsformanufacturing.com/departments/logistics/)
- Evidence level: Verified. Real deployment at a named or precisely described company, with a measured result from a primary source.
- Added: 2026-10-30
- Last updated: 2026-10-30
- Publisher: AI Agents for Manufacturing (https://aiagentsformanufacturing.com)
- Manufacturers named: HiKOKI (https://aiagentsformanufacturing.com/companies/#hikoki), Metso (https://aiagentsformanufacturing.com/companies/#metso), Mitsubishi Electric (https://aiagentsformanufacturing.com/companies/#mitsubishi-electric)
- AI tools deployed: PTC Servigistics (https://aiagentsformanufacturing.com/tools/#ptc-servigistics), ToolsGroup SO99+ (https://aiagentsformanufacturing.com/tools/#toolsgroup-so99)

## Role

After-sales and service parts supply chain (spare parts sold to customers, dealers and service centers). Typical users: service parts planner, after-sales or spare parts manager, supply chain manager, warehouse manager

## The problem

Manufacturers that sell machines, equipment or tools must keep spare parts for customers for years. Demand for each part is low and irregular, there are thousands to millions of part-location combinations, supplier lead times are long and customers expect parts within hours or days. Standard ERP min/max settings and history-based planning then pile up obsolete stock in some places while critical parts are missing elsewhere. At Mitsubishi Electric Europe, one spare parts business of its heating, cooling and ventilation division was nearly doubling obsolete stock every two years to reach an 87% service level; demand is 'hyper-seasonal' because it follows the weather, and most repairs are urgent. At HiKOKI (formerly Hitachi Power Tools), the European service network had 22 locations and over 70,000 part numbers, or 1.6 million part-location combinations; an analysis found more than 60% of stock was dead or in the wrong place, while the market expects delivery in 48 hours and some supplier lead times approach six months. At Metso (mining and process equipment), 39 parts centers each managed their inventory in isolation.

## How AI helps

Specialised service parts planning software forecasts the irregular demand of each part at each location (statistical and machine learning methods built for slow, intermittent demand, with seasonality detection), then sets the stock level each part needs at each location to reach the service level chosen for it, looking at the whole network rather than each warehouse alone (multi-echelon optimisation). Planners get alerts instead of checking every part: demand that is higher or lower than forecast, stock above or below its range, deliveries to postpone, and transfers between depots. The plan goes back to the ERP as purchase requests or stock transfers. Mitsubishi Electric Europe's ToolsGroup system gives each part a stock 'corridor', detects seasonality automatically, can forecast new parts that replace old ones, and is described as self-learning. HiKOKI and Metso use PTC Servigistics; their case studies describe the planning and optimisation methods but do not say which parts use AI.

## Step by step

1. Measure the starting point: service level (share of orders shipped complete on time), backorders, total parts inventory, and stock with no movement in 12 or 24 months, by location.
2. Segment parts by criticality for customers and by demand pattern (regular, seasonal, slow or erratic, new or being phased out), and set a target service level for each segment instead of one figure for all.
3. Clean the master data that planning depends on: lead times, supersessions (old part replaced by new), minimum order quantities, and the installed base of machines still in service if you have it.
4. Pilot on one product family or region: run the tool's recommendations in parallel with your current settings and compare, part by part, where they differ and why.
5. Expect inventory to rise for some parts and fall for others while the mix is rebalanced (Metso's total stock rose for four months before falling); agree this with finance before starting.
6. Work by exception: planners review alerts and transfers, not every part, and track service level, backorders, inventory and dead stock monthly.

## Tools

- ToolsGroup SO99+ (demand forecasting and inventory optimisation, integrated with SAP): used at Mitsubishi Electric Europe
- PTC Servigistics Service Parts Management (with multi-echelon optimisation at Metso): used at HiKOKI and Metso
- Other tools to evaluate: Syncron, Baxter Planning, Lokad, the service parts planning module of your ERP (for example SAP)
- For a first step without new software: a parts and demand extract from your ERP in a spreadsheet, plus any LLM (ChatGPT, Claude, Copilot, Gemini) with the prompt below

## Result

Mitsubishi Electric Europe (heating, cooling and ventilation spare parts), ToolsGroup SO99+ integrated with SAP, live five months after a pilot: spare parts stock down 30% and service level up from 87% to 97% within three years, even during seasonal peaks (ToolsGroup customer story, a vendor claim; quote from Thomas Schumann, Mitsubishi Electric Europe). HiKOKI (power tools, European service network), PTC Servigistics: inventory down 34% across the network, dead or wrong stock in distribution centers down 65%, backorders from four months to four weeks, while central stock rose 13%; two stock planners at the central warehouse manage the network by rules and exceptions (PTC case study, undated, a vendor claim). Metso (mining and process equipment, 39 parts centers), PTC Servigistics, live in June 2014: inventory rose for the first four months while the part mix was rebalanced, then fell; more than €30 million saved in the first phase and €11.6 million more after adding multi-echelon optimisation in October 2015, €41.6 million in total, inventory turns up 18% in 24 months and on-shelf availability up 3.4% (PTC case study, a vendor claim).

As reported by the source; most figures are vendor or self-reported and not independently verified.

## Human review: cautions

- All three results are published by the software vendors (ToolsGroup, PTC); the Metso project dates from 2014 to 2016 and the HiKOKI case study is undated. They cover a whole planning change (process, data, software), not AI alone.
- Service level and inventory can always be traded against each other: check that a lower inventory did not come with more backorders, and the other way round.
- Planning is only as good as lead times, supersessions and demand history; a few wrong lead times or missing part replacements can create both shortages and dead stock.
- Scrapping or moving dead stock has accounting effects; involve finance before writing off parts identified by the tool.

## Try it yourself (prompt)

```text
You are a service parts planner at a manufacturer that sells spare parts to customers and dealers.

I will paste an ERP extract for one product family or warehouse [part number, description, location, on-hand quantity, unit cost, monthly demand for the last 24 months, supplier lead time, current min and max or reorder point, superseded by, date of last movement].

Do this:
1. Classify each part by demand pattern (regular, seasonal, slow or erratic, no demand in 12 months, new) and explain the rule you used; tell me which spreadsheet formulas to use for the averages and variability, and do not compute them yourself.
2. List parts with stock but no movement in 12 and 24 months, and parts that are superseded but still stocked, copying the quantities and costs I gave.
3. List parts whose current min or reorder point looks too low or too high for their demand and lead time, with one line of reasoning each, labelled as a hypothesis to check.
4. Propose 3 to 4 part segments with a suggested target service level for each, based on criticality and demand pattern, and the questions I need to answer to confirm them.
5. Draft a one-page plan for a 3-month pilot of a service parts planning tool on this family: what to measure before, during and after.

Use only the data I paste in; if something is missing, write 'not found' or list it as a question instead of guessing. Do not send, order, post or change anything in any system: produce drafts for my review and approval.
```

## Related use cases

- [AI replenishment recommendations so reorders are not missed](https://aiagentsformanufacturing.com/use-cases/ai-replenishment-recommendations/)
- [AI finds duplicate and excess MRO spare parts across plants](https://aiagentsformanufacturing.com/use-cases/mro-spare-parts-inventory-optimization-verusen/)
- [Technicians identify a spare part from a photo or a few words and see its stock and location (Deutsche Bahn, ENGEL)](https://aiagentsformanufacturing.com/use-cases/ai-spare-part-identification-photo-search-partium/)
- [AI moves fast-selling items closer to the pickers and shortens pick routes in the warehouse (Coram, Gildemeister)](https://aiagentsformanufacturing.com/use-cases/ai-warehouse-slotting-pick-path-optimization-coram-gildemeister/)

## Sources

- https://www.toolsgroup.com/customer-stories/mitsubishi-electric/
- https://www.ptc.com/en/case-studies/hikoki
- https://www.ptc.com/en/case-studies/metso-spm-case-study
- https://www.toolsgroup.com/blog/how-ai-is-transforming-aftermarket-parts-planning/

---
Please cite the canonical URL: https://aiagentsformanufacturing.com/use-cases/ai-service-parts-inventory-planning-aftermarket-mitsubishi-electric-metso/