UNIMAT Retirement Community

This system was jointly developed with Macnica. Macnica, since it acquired 41.8% of share of CroudANALYTIX, a start-up company in India, in January 2019, has accelerated its AI business. In the UNIMAT RC’s AI system, we used
the AI platform services possessed by CrowdANALYTIX.

With this service, we can simultaneously develop various AI modules by utilizing more than 20,000 data scientists worldwide. Specifically, we disclose the bare minimum specifications required to develop a module that coincides
with the needs of the client, simultaneously develop an algorithm worldwide, pick one that excels in accuracy, and then develop a module which combines them.

So we incorporate what they call a “community of data scientists” into our system development. As a side note, data scientists are engineers who are active in companies worldwide, such as leading platformers.

Mr. Okada says, “When you introduce AI, an accuracy of 80% is insufficient. In that sense, CrowdANALYTIX’s AI platform service was created by a specialist group with excellent ideas. Moreover, a system that allows these other
data scientists to assist at any time is just wonderful.” Mr. Takeo Nishimura, AI Business Promotion Director of New Business who promotes AI businesses at Macnica, states, “Development is done by communicating with Japanese engineers but
the fine tuning work is done by CrowdANALYTIX.”

As mentioned above, we initiated a demonstration trial on streamlining a vacancy reservation management through the use of AI from August 2019. Reservation management at short-stay facilities, unlike hotel reservation
management, requires that a wide variety of factors be taken into consideration. “Those factors include things like the preferred orientation of the bed depending on which side of the body is paralyzed, preferred location being close to the
bathroom for reasons attributed to medication with a diuretic effect, or preferred location being near the staff observation room for reasons attributed to a risk of falling. Further, admission and discharge hours may need to be adjusted
even if the length of stay is the same, which makes the work almost like a puzzle.” Says Mr. Okada. Room assignment can require quite good skill.

Generally, counseling staff takes care of reservation managements using hand-written tables and Excel, but their skills vary depending on the person, leading to occasional inefficient room assignment. Sometimes, simply
switching rooms solves the problem. At UNIMAT RC, we decided to increase efficiency by equalizing the work through AI in response to these challenges. “When we had the AI run the numbers based on historical data, more efficient room
assignment was instantly created. There are high expectations for it to contribute to the improvement in labor productivity of counseling staff,” says Mr. Okada. Even when involving simply switching rooms, “this contributes to the reduction
in the emotional strain placed on counseling staff, who have to make adjustments taking various factors into consideration. We can assign rooms with minimal room switching, thereby reducing stress on the residents as well,” says Mr. Okada
with great expectation.

Demonstration trials of AI applied at the two short-stay facilities will begin full operation by the end of 2019. Thereafter, we will deploy this at around 140 short-stay facility locations by the end of 2020 and plan to
propose this at other nursing homes jointly with Macnica. “As for usage fees, we are considering methods such as one in which we get paid for the amount of cost reduction achieved, one in which we get paid for the operational contributions
made, or a recurring model in which the fees fluctuate according to usage,” says Mr. Nishimura (Macnica). Both companies plan to utilize AI in the field of nursing care, in which the introduction of IT is slowly taking place and support
their operational streamlining.