Hello! This is Fukuoka from Macnica 's Digital Industry Division.
I went to JIMTOF2026 (The 33rd Japan International Machine Tool Fair), which was held at Tokyo Big Sight from October 26th (Monday) to 31st (Saturday), 2026. It is a major trade fair for machine tools, with 1,134 exhibitors from 22 countries and regions around the world.
This time, we'll focus on the use of digital technology in after-sales service/customer success (CS) areas by equipment manufacturers, and share what we learned at each company's booth.
Overall technology trends
To sum it up in one sentence: "AI agents are everywhere." As I walked around the venue, I encountered the same words over and over again: AI agent, AI chat, CX... To summarize, there are three main points.
Last time, the customer portals and inquiry handling displays were rather understated, but this time, major companies were all putting them front and center.
By feeding manuals, instruction manuals, and parts lists into a generating AI, we were able to create inquiry response and knowledge search services in a relatively short period of time. It's also impressive that many of these services are developed in-house.
In the realm of general-purpose precision, Chinese and Taiwanese manufacturers are catching up, making it difficult to differentiate based solely on the specifications of the machine itself. The challenge now lies in how to increase the added value of the machine through digital technology.
On the other hand, there are still few services that have already launched, with most being showcased as reference exhibits or scheduled for release next year or later (as far as I could see, DMG Mori Seiki was the only company that had released a CS service using generative AI, and Yamazaki Mazak was also planning to start offering it around mid-November). There seem to be challenges to overcome before reaching"operational quality," such as variations and inconsistencies in response quality. I got the impression that each company was promoting their adherence to technology trends while simultaneously marketing "how to sell" at JIMTOF.
The AI corner was packed with people everywhere (Okuma's booth).
Transforming support with AI agents
Yamazaki Mazak
Yamazaki Mazak always impresses visitors with their enormous booths. One of the highlights this time is the AI agent integrated into the inquiry form of their customer cloud service, "Mazak iCONNECT".
Previously, the process involved filling out an inquiry form, sending an email to the service support center, and then receiving a follow-up email or phone call. Now, an AI agent is integrated into the form, providing an immediate initial response. The training data includes manuals, inquiry information received at the call center, and reports from service engineers performing repairs at customer sites. Its strength lies in the fact that it incorporates the field knowledge accumulated over many years.
The key is how you handle situations where the problem isn't solved. If you reply, "I replaced the battery, but it's still not fixed," they'll guide you to another solution. If that still doesn't work, you can contact the call center via chat, and since the chat history is carried over, the call center staff can guide you to the next step without having to ask, "Have you tried this yet?"
The presenter, who previously worked in a call center, explained that this initiative stemmed from his experience of feeling sorry for customers having to wait an hour before being able to return calls. The company's service technicians are already using it, and it will be offered to customers free of charge through the "Mazak iCONNECT"web service around mid-November. The exhibit directly addressed customers' desire to "get their machines running as quickly as possible," even delving into the customer service workflow.
On the PC application side, there was a demonstration of the subscription-based software "MAZATROL DX," where an AI agent utilizes operation status and machining data to support everything from program creation and setup to troubleshooting. It can also suggest cutting conditions for materials that have never been used before and respond to voice inquiries.
This is the inquiry screen for Mazak iCONNECT. An AI agent has been added here.
Mazak iCONNECT and MAZATROL DX both feature AI agents.
DMG Mori Seiki
At DMG Mori Seiki's booth, which prominently featured "CUSTOMER CARE," they showcased "Service Agent," an AI usable on their customer portal "my DMG MORI." Released just about a month before JIMTOF, it's one of the few AI-generated CS services that is "already usable" as far as I could see.
The goal is to make service inquiries understandable in a single step. To reduce the"ping-pong"situation—where the situation isn't clear from the inquiry alone, leading to repeated callbacks and back-and-forth with the customer—AI helps create detailed requests that include all the information necessary for the service to respond. If a solution is available in the manual, it will point you to the relevant section, allowing customers to troubleshoot problems themselves.
The knowledge base currently only contains manuals for customers. Since it's not connected to the machines, it can't acquire operational status data, but this is being considered for future development. It's offered as part of the my DMG MORI service and there are no additional charges. The "Condition Agent," which will be standard on the new models, was also introduced.
When I asked, "Is it difficult to beat the competition without strengthening after-sales service?", the representative replied, "I think after-sales service is the most important thing." It was a statement that perfectly encapsulated this year's JIMTOF.
A corner dedicated to "Customer Care"
AI-powered support for inquiries: "Service Agent"
Brother Industries, Ltd.
Brother Industries showcased a "Web service with AI chat support for SPEEDIO operation" as a reference exhibit. Customers who purchase the machine can be issued an account and, upon logging in, can do the following:
- Ask AI chat questions about part numbers, programming methods, and parameter settings.
- Check the specifications of the machines you own and their past repair history.
- Reference to processing examples, e-learning (paid plan)
- Real-time operational monitoring of machines equipped with communication units (expected to be a paid service)
The source data for the AI chat is from instruction manuals and parts lists, and it's developed in-house. They mentioned that they're still working on creating customer-specific knowledge so that it "gets smarter for each customer." You can also try it out on your smartphone by scanning a QR code at the booth, and before you know it, you've become a user yourself.
Web service with AI chat to support SPEEDIO operation (reference exhibit)
The program consists of two parts: "Getting to Know SPEEDIO" and "Managing SPEEDIO."
Sodick
Sodick presented its "Sodick CS Solutions to Support Uninterrupted Factory Operations," offering a package of after-sales services including extended warranties, regular inspections, and parts replacement packages, along with its integrated remote service, "SatelinC." Their presentation, showcasing services as"products,"is a valuable reference for anyone considering a CS business.
SatelinC allows you to check the status of your machinery remotely using a smartphone or other device, and automatically compiles reports on operating rates and downtime causes. It's a paid service with an initial fee plus a monthly fee, and they've already received inquiries from several companies.
In addition, they showcased an AI solution that learns from instruction manuals and customer support manuals, and then answers questions about the meaning of errors and how to resolve them in a chat format. They mentioned that it was still in the prototype stage and not yet ready for commercialization.
After-sales service menu and integrated remote service "SatelinC"
Check the machine status from your smartphone
Shibaura Machine Co., Ltd.
Shibaura Machine Co., Ltd. is strong in ultra-large machine tools, such as gantry-type machining centers resembling shrine gates. In addition to exhibiting "machiNetCloud," which connects machines using individual protocols and the secure standard OPC UA, they introduced "LEO," an AI assistant that answers questions by incorporating manuals and machine information into a knowledge base. They plan to release it in the US first in the spring of 2027.
Because it requires a cloud connection, there are still high hurdles to overcome in Japan. Currently, there is no function to refer to each customer's trouble history, which will likely be a challenge for the future.
"Hi, I'm LEO." (AI assistant screen)
machiNetCloud connects machines using OPC UA.
Creating programs using natural language
FANUC
FANUC, a supplier of CNC systems to many machine tool manufacturers, is promoting "Physical AI CNC," and the highlight of this project is the MCP server. Until now, AI chat and various software were not directly connected, but by using the MCP server, it is now possible to call tools from the AI chat.
In the demo, the AI refined a hand-drawn image on a tablet, called up a tool to convert it into G-code, verified the G-code with CNC Guide 2, and confirmed the finished product with a machining simulation—all in a seamless process via chat. The idea is that AI can now operate tools that were previously operated by humans.
I drew a cat on my tablet...
The AI calls up a tool and says, "Testing operation with CNC Guide 2."
Incidentally, when I asked them to "make the cat illustration stylish," they created an illustration of a cat with brass instruments as its motif. Expressing style through jazz is quite a clever touch. As a souvenir, I received a plate that they had processed on the spot.
The finished product. The jazz cat is quite stylish.
Another feature is interactive machining program creation. By specifying 3D CAD data and saying "the material is 92mm in diameter and 150mm in length," the material is set, and once the work coordinate system is determined, the necessary processes are automatically generated. If you say "please process the outer shape first," it will also rearrange the order of the processes. It is scheduled for release at the end of August 2027.
When I asked if the generated program was truly correct, they said, "There's no problem." It seems the AI 's role is to listen to instructions and guide the operation, while the program itself is created using a fixed logic. I felt this was a good compromise for using generative AI, which tends to produce inconsistent answers each time, in an industrial setting.
A program instructed in natural language is simulated on the spot.
Mr. Okuma
Okuma introduced their "two digital twins" system, where they simulate on-site machinery on an office PC to check for interference and estimate processing time before actually performing the work on the machine. They emphasized the ease with which even beginners can create programs using "Smart OSP operation." According to the demonstrator, the precision is comparable to that of products made by skilled craftsmen.
The AI corner was quite crowded, a scene that truly symbolized this year's trends.
Okuma's two digital twins
Even beginners can easily create programs with "Smart OSP Operation".
It's not just about AI, it's about grounded digital transformation.
Makino Milling Machine Co., Ltd.
Makino Milling Machine Co., Ltd. exhibited its production support software suite, "Makino Advanced Software." This time, rather than focusing on generative AI, they emphasized strengthening the fundamentals of automation and monitoring.
The scheduling function of the automated system, which combines processing machines, setup stations, and pallets, automatically creates a processing sequence using a rule-based algorithm based on inputted deadlines and priorities. If there is a deviation from the actual schedule due to alarm stoppages or other reasons, the plan is readjusted once processing is complete.
Makino Advanced Software, a suite of production support software.
A scheduling screen that automatically creates the processing order based on delivery date and priority.
Komatsu NTC
Komatsu NTC, which handles entire mass production lines for automobile engine parts and other components, showcased its machine tool monitoring system "KOMTAS" for detecting tool wear and chipping. The system captures data through detailed sampling that reveals the work of each individual blade, and then scores the data after removing unnecessary parts such as air cuts.
Regarding wear, tools that were previously replaced after a certain number of uses based on experience can now be used to their full lifespan by observing their actual condition. Chipping (breakage of the cutting edge) can be treated as a workpiece requiring inspection, preventing defective products from proceeding to the next process. The system is designed with the user's perspective in mind, allowing for different stopping methods, such as "not proceeding to the next processing step" if there are signs of tool breakage. It is already available as an option.
Because it's a mass production line that repeats the same processing every day, anomalies can be identified by comparing yesterday's and today's data. I felt this was an approach unique to a mass production machine manufacturer.
"Making anomaly monitoring easier" -KOMTAS machine tool monitoring system
Nakamura Tome Precision Industries Co., Ltd.
Nakamura Tome Precision Industry Co., Ltd. showcased the automatic threshold setting function of their tool breakage and wear detection solution, "Dr. Tool," in the "Bridg3" corner. After the first five or more machining operations, the system automatically sets the threshold based on the results.
When I inquired about their efforts with generative AI, they said it was still in the product planning stage. "Industrial products require repeatability and accuracy. It would be problematic if the answers were different every time," was perhaps the most impactful comment I heard during this visit.
Bridg3 corner. "Operator-free, preventative maintenance operation becomes possible."
Citizen Machinery Co., Ltd.
Amidst a sea of chatbots, Citizen Machinery's estimation support/drawing management tool stood out. When you upload a drawing, the AI searches past drawings in order of similarity, and you can check the associated estimation history and process. It also shows the differences, so anyone can come up with a draft that's about 70 % complete by referring to past estimates.
It's not uncommon to hear from equipment manufacturers that "even though I'm an engineer, 40 % of my work is spent on quotation preparation." It could also be used in design departments to prevent situations where "a neighboring department was making a similar part, but we had to redesign it from scratch."
Quotation support/drawing management tool
Compare similar drawings with past estimates.
Star Micronics
Star Micronics exhibited its conversational program creation support software, "PU Advance." Claiming to "support the standardization of NC programming operations," it allows users to create NC programs from scratch simply by inputting the necessary items. It also supports input completion, calculation of cutting conditions for each material, and 2D display to check tool paths before machining.
In addition, it allows users to manage programs, drawings (DXF / PDF), and photos of parts all on a single screen, reducing the effort required to search for past programs when placing repeat orders. The price is a one-time purchase of 150,000 yen in USB dongle format, and a 90-day free trial is also available.
When I mentioned that it would be even more convenient if the AI could automatically find similar past programs and even create programs based on them, they said they plan to add various features in the future. I'm looking forward to seeing how it evolves at JIMTOF in two years.
Conversational program creation support software "PU Advance"
Manage drawings, photos, and programs on a single screen.
Summary: The key to success going forward will be "operational quality".
At this year's JIMTOF, it became clear that CS services using generative AI have entered the stage where they can be "created." It was also striking to see the stark contrast between companies that are working on AI agent development and those that are not, reflecting the democratization of AI agent development.
On the other hand, from what I've seen, DMG Mori Seiki is the only company that has already released a CS service using generation AI (Yamazaki Mazak plans to start offering it around mid-November), and many exhibits are for reference only or scheduled for release next year or later. Challenges are also becoming apparent.
- Variation and inconsistency in responses: In the field of industrial products, it is assumed that the same answer will be given every time.
- Customer-specific knowledge: Beyond simply reading manuals, how to leverage each customer's specific machine, history, and troubleshooting experiences.
- Business Model: Selling AI chat as a standalone product is difficult. One possible solution is to integrate it with improvements to your company's service operations and expand customer touchpoints.
While the AI provides an initial response, the service technician can consider the next step. The customer gets an answer immediately. Both the person asking and the person answering feel good—as the differences in equipment specifications narrow, I felt that how you design the "after-sales experience" will become the key to the competitiveness of machine tool manufacturers going forward.
At JIMTOF in two years, we should see improvements in operational quality, successful implementation examples in actual operations, and how the system has been firmly integrated into our customers' business processes. I am already looking forward to seeing how the manufacturing industry will further evolve through digitalization.
See you in two years at JIMTOF2028!
That concludes our presentation. Macnica supports equipment manufacturers with after-sales service DX and the construction of data services in the customer satisfaction (CS) domain. If you are interested, please feel free to contact us!
*The content of this article is based on exhibits and presentations given by staff at JIMTOF2026 in October 2026. It may differ from the official announcements made by each company.