Supporting Social Impact Through Responsible AI
Photo: Temus’ AI & Data Team with NCSS organisers at NCSS Service Summit 2025
At the 2025 NCSS Service Summit, Temus participated in TribeX to demonstrate how technology — when built with intention and empathy — can help deliver real social impact. Our showcase, DialogueForge, represents this commitment: a conversational AI solution purpose-built to support the meaningful work of Social Service Agencies (SSAs) in Singapore.
Our presence aligned with Minister for Social and Family Development Masagos Zulkifli’s call for an “IDEAL” sector — one that is Innovative, Data-powered, Efficient, Accessible, and Linked. His emphasis on building tech with real social impact deeply resonates with our mission: to bring responsible, human-centred innovation to public good.
We believe that AI, when thoughtfully applied, can extend the reach and empathy of social service agencies, not replace it.
How DialogueForge Works: Real-Time Conversations, Real Support
Powered by a chained large language model (LLM) architecture, DialogueForge enables ultra-low latency, real-time speech-to-speech interactions through a conversational AI avatar. In practice, this means individuals can speak naturally to an animated digital companion, which listens, processes their needs using contextual understanding, and responds — all within a few seconds.
As demonstrated in our video walkthrough, the avatar initiates natural dialogue almost instantly, simulating a human-like conversation to guide users through assessments, answer questions, or recommend next steps. It’s designed for high engagement, high empathy, and high scalability — tailored especially for public sector and nonprofit use.
Meet Kai: A Friendly e-Buddy for Youth Engagement
One early example is Kai, an AI-powered companion we designed to support youth outreach. Kai listens, learns, and helps assess early needs — making it easier for young individuals to access help or information, while supporting case officers and social workers in their outreach efforts.
Following a conversation with Kai, the eBuddy, a comprehensive session report can be automatically generated to assess the user’s overall wellbeing. This report goes beyond surface-level responses – analysing engagement levels, identifying emotional triggers (such as references to smoking), and interpreting the user’s current emotional state.
Through linguistic analysis, the system picks up on subtle language cues, such as the use of short or dismissive sentences, tense shifts (from past to present), and other psychological patterns that may signal distress or disengagement. These insights are used to provide an initial risk assessment, ensuring that any red flags related to the user’s mental or emotional health are promptly detected.
Based on the findings, Kai can also suggest personalized wellbeing tips, propose follow-up sessions, or flag the case to a human professional for timely intervention – supporting SSas in delivering proactive, data-informed care with empathy at scale.
Learn more about our AI & Data Services
We’re committed to helping SSAs and mission-driven organisations harness responsible AI to deepen their impact.
Discover how Temus’ AI and Data services can support your mission.
