Skip to content
Interface language: English
Oriens / Signal

Service robots in Japan: a demonstration is not yet the machine’s workplace

Investigate the task, facility conditions and roles of users, managers and integrators before researching service robots in Japan.

Article language: English

A person wearing a glove touches a red flexible-looking material.
In this article

At a demonstration, a robot travels across an empty hall. In a real building, a cart crosses its route, someone moves a chair and a door closes through a system absent from the video. That is a particular service-robot task in a particular environment, and a useful starting point for market research. A Polish company seeking an application for its technology in Japan needs to describe that environment before counting every hotel, office building and shopping centre.

Start with cleaning or moving objects in a non-medical facility. Who commissions the work, prepares the space and responds when the device stops? The answers help identify the application and commercial relationship worth investigating.

Vacuuming is a task; a building is a set of conditions

SoftBank Robotics' Whiz series page distinguishes models and operating methods intended for different floor conditions. For Whiz and Whiz i it describes teaching a route by manually guiding the machine. This is a manufacturer's description of a feature of specified models, not proof of a result in every building.

That detail changes the research questions. If a robot uses a prepared route, establish how often the space changes and who can alter it. If it must perform a particular kind of cleaning, floor area alone does not describe the task. The zones, their use and work remaining for staff matter. Multiplying a building's entire area by catalogue performance does not make an operating plan.

In a fictional office building, reception is organised differently from a floor with movable partitions or an event room. The administrator wants to reduce a cleaner's repetitive trips, but still needs the area prepared, containers emptied and the result checked. At this stage we do not assume the robot performs all those tasks. First record the existing workflow and select a part to evaluate. Listing work it does not perform helps define an application without promising to replace the entire service. If the stated process does not explain staff's real work, agree open observation of that work before deciding on a change (in Polish), within rules appropriate to the facility.

A purchasing conversation might therefore begin, "we need a way to maintain this zone at specified hours", rather than "we need a robot". That description allows comparison of configurations and asks whether a Polish company supplies the needed element: a device, software, environmental perception or a specialist service. Not every supplier must offer a complete system, but each should state the boundary of its contribution.

For B2B market segmentation (in Polish), distinguish facilities by task and constraints. A hotel whose shared area is regularly rearranged may need different arrangements from a repetitive office corridor. These are criteria to investigate, not a claim that one type of building is always easier. Industry name gives an initial filter; a description of the workplace supports a conversation about fit.

Before a first demonstration, describe an ordinary day. When is the zone open, when does its layout change and when do other activities take place? In the fictional building the same room serves meetings in the morning and events in the evening. A photograph of the architecture is not enough for a robot; describe recurring changes that may affect preparation.

Record the beginning and end of work too. Who prepares the device, confirms the area is ready and checks the result? If an external company currently performs those activities, an administrator should not assign them to its staff without agreement. Buying a solution can change service organisation. The technology discussion must therefore meet the discussion of people's work, rather than assume new tasks will find owners by themselves.

A nearby lift is not automatically a robot feature

The City of Sagamihara's notice of 3 March 2026 described a trial in which a robot travelled between floors of a municipal building using a lift. The city announced further improvement of functions and operating rules after evaluation and surveys. This is an example of a demonstration stage, not a universally ready standard for all buildings.

A company researching the market needs to consider connection to another party's infrastructure. The robot, door system, lift and access arrangements may belong to different parties with different responsibilities. Technical arrangements require the right people, but a researcher should already understand that "a building interested in a robot" is not one consent covering the device's entire route.

Imagine a service transporting closed containers between a back room and a collection point in an office complex. This is hypothetical. On a short route the robot may work without changing floors. Expanding the route introduces doors, access boundaries and a lift, changing the discussion's scope. A successful test on one floor does not automatically establish capability on all floors.

Show route segments and dependencies in the research material. You need not design an integration to identify a need to check it. "Confirm cooperation with the access system" is more useful than a generic "technical risk": it identifies a question, the right person and why the scenario cannot yet be expanded. It also avoids seeking demand for a solution the company has not prepared for the proposed use.

User, facility manager and service provider may buy different things

A building owner may consider long-term organisation of space. A manager must maintain everyday operation. A cleaning company considers changes to its own service. A technical team asks about infrastructure and maintenance. Do not assume one role always approves the purchase. A map of B2B purchasing participants in Asia (in Polish) helps assign questions to people in the chosen case.

Aeneia conceptual photography

For a Polish obstacle-detection module maker, a device supplier or integrator, an organisation combining elements into a working solution, may be the right partner. A visit to a hotel administrator need not immediately be a purchasing conversation. It may provide knowledge of a task to address later with a suitable partner. This distinction avoids mistaking user interest for a ready route to an order.

A similar issue appears in the Korean smart-factory ecosystem (in Polish): the end facility, component supplier and system implementer have different roles. The comparison does not prove identical channels in Japan and Korea. It helps ask where the offered solution sits in the chain of use and who can assess its usefulness.

When recording contacts, distinguish confirmed roles from hypotheses. An innovation lead may invite a demonstration without managing cleaning. An installation contractor may know infrastructure but not be responsible for staff's service scope. Record these differences before a presentation so you do not expect a decision outside someone's role. A useful meeting does not require treating every participant as the head of the entire building.

Separately check how the company describes service after delivery. Support may mean operator advice, remote analysis of a report or a specific action on site. These are not interchangeable promises. Ask a potential partner about one ordinary case: the device has stopped, staff do not know why and the task still needs completing. Who receives the information, who can help and what does the facility do until the matter is resolved?

The answer need not guarantee no stoppages. It should show a real course of action. If the Polish company supplies a component, some information may reach it through an integrator while other information stays with the device operator. Research should establish the relationship, not invent an unagreed promise of direct support for every user. This matters especially when sales material says "support" without saying who actually receives it.

A trial should include a day when something gets in the way

A device demonstration and a trial of work organisation answer different questions. The former can examine a specified feature under prepared conditions. The latter must show how a team uses the solution in its environment. Supplement a sample and demonstration evaluation plan (in Polish) with activities before start, during work and after completion. Otherwise it is easy to measure travel time alone.

In our office building record three kinds of observation: whether the designated zone was served, how many interventions were needed and what staff had to do. We give no universal thresholds; they should match the task and agreed assessment. An intervention is not automatically a failure. It is an event to understand and include in actual service organisation.

Also observe layout changes, a blocked section and handover between shifts. Appropriate specialists responsible for the robot and facility must assess device safety, technical compatibility and trial conditions. The point is not to provoke dangerous situations yourself. It is to establish what confirmation is needed before promising everyday operation. A video of a successful run does not tell staff how to finish work after a stop.

In the office example distinguish a robot failing to cover part of the route from one completing the route while the zone is not ready for use. The first raises questions about movement and surroundings. The second may concern the task's scope, evaluation method or work left to staff. Combining both as "route completed" hides a distinction that matters to the manager. The service recipient wants to know the task is done, not merely that the device reached the end of its recorded route.

Describe fallback with equal care. Which work returns to the previous method, and who decides? This is not a claim that the robot will fail often. Every service needs a way to finish when a tool is unavailable. If a supplier representative helps during a demonstration, record their actions. Then establish whether ordinary operation assigns them to facility staff, a service partner or another named party. This reveals the real scope without adding an invisible helper's work to the product.

After the trial compare planned and actual scope. The device may be useful in one zone while others need extra work or lie outside scope. That need not rule it out; it may lead to an honestly narrower offer. The problem would be presenting a result from the easier part as proof the entire facility is ready, especially while staff still perform omitted tasks.

When expanding a trial, do not change every condition at once. If another facility has a different floor, traffic pattern and operator, similar area alone does not justify expecting the same outcome. State what is being compared and which differences require checking. The next person can then use the application finding without inheriting an unconfirmed result.

Where does the tested application fit in the offer?

After investigating an application, a Polish company can choose a narrower proposal: prepare material for integrators rather than advertise complete building cleaning, for example, or focus on a particular space before describing an offer for all commercial buildings. That decision should follow findings about the task, available partners and its own product scope, not merely the number of visitors to a robotics fair.

Japan's machine-tool industry (in Polish) illustrates a similar need to distinguish equipment from services for existing infrastructure. Machine servicing in Poland (in Polish) offers a view of events after delivery. In both, technology usefulness extends beyond a feature demonstration. These are comparisons of tasks, not transfers of technical conditions between different machines.

At Aeneia Oriens, research can organise applications, stakeholder roles and material for an initial assessment. It should yield a question suitable for a concrete discussion: with whom should a particular part be checked and which conditions need description first? A robot needs more than general admiration for technology in Japan. It needs a task, a workplace and people who know what to do after the demonstration.

Aeneia / Aeneia editorial

Editorial article. Source links are included in the text.

About Aeneia

Brands we have worked with

  • L’Oréal
  • Garnier
  • Netflix
  • Carrefour
  • Warner Bros. Discovery
  • Biedronka
  • HBO Max
  • Wolt
  • Żabka
  • COSRX
  • Embassy of China in Poland
  • Mixa
Aeneia / Initium

From knowledge to a decision

Tell us what you want to put in order. We will agree on where to start.

Contact Aeneia

We partner with and use solutions from:

  • Google
  • OpenAI
  • Microsoft
  • Stripe
  • Anthropic
  • Linear