Page 90 - AI for Good Innovate for Impact

AI for Good Innovate for Impact

(continued)

Item Details
Testbeds or Pilot The solution has undergone extensive field validation through multi-re-
Deployments gional pilot deployments, ensuring real-world usability and scalability.
• Gimhae City Deployment: AI Senior Care Content has been actively
implemented in 36 smart senior centres, with plans to scale the
deployment to over 50 facilities nationwide. These centres serve as
a core testbed for evaluating content usability, interaction patterns,
and health-related engagement metrics.
• National Pilot with Ministry of Health and Welfare: A large-scale
demonstration involving 500 elderly participants is being conducted
under the supervision of the National Rehabilitation Centre. The
project tests the effectiveness of the AI Conversational Companion
Service across various user profiles and care environments.
• International PoC with Japan (SpiralAI) [7]: Collaborative Proof-of-Con-
cept projects are being carried out with SpiralAI and local Japanese
governments. These pilots are focused on validating cultural adapt-
ability, multilingual LLM performance and integration into Japan’s
ageing care ecosystem.

Code repositories Not publicly disclosed

2 Use Case Description

2�1 Description

In response to the escalating challenges posed by ageing populations, ESTsoft has developed
a comprehensive AI-driven solution to enhance the emotional, cognitive, and physical well-
being of seniors [3]. This initiative comprises two core components: AI Senior Care Content and
the AI Conversational Companion Service. The content offering includes over 180 specialised
programs, such as traditional music, guided exercise, and cognitive training games, currently
deployed across senior centres via television-based platforms. A key strength of the service lies
in its advanced personalisation capability. Each content module is systematically tagged using
a health classification framework adopted from the Ministry of Health and Welfare [5] and the
National Rehabilitation Centre [6]. When a senior interacts with the AI avatar, their inputs and
behavioural patterns are processed through a classification algorithm that identifies and delivers
the most suitable content. As usage data accumulates, ESTsoft aims to evolve this system further
using machine learning, ultimately enabling dynamic, data-driven personalisation at scale [1].

Building on this personalised approach, the AI Conversational Companion Service delivers real-
time, lifelike interactions through AI avatars. These avatars are equipped with core functionalities
including health monitoring, memory assistance, and dementia prevention, all underpinned by
a senior-optimised Large Language Model (LLM) and high-precision Speech-to-Text (STT) and
Text-to-Speech (TTS) technologies. Real-time feedback is continuously gathered from users
across 36 senior centres and partner medical institutions. Integration with wearable devices,
such as smartwatches, enables seamless tracking of key health indicators, including heart
rate, sleep cycles, and physical activity. These data streams form a robust feedback loop that
enhances service quality, facilitates adaptive content delivery, and ensures timely, individualised
support for seniors.

54