The useful question is what changes for users, developers or buyers, and whether the announcement stays industry context or becomes something people can actually use.

Siow Shong Seng, chief transformation and technology officer at Singapore National Trades Union Congress (SNTUC), has deployed Microsoft 365 Copilot to automate routine tasks such as generating daily updates on labor trends and social media monitoring. He built an AI agent to assist industrial relations officers in retrieving labor laws and reviewing employment contracts, drawing from sources including SNTUC and the Ministry of Manpower. The agent serves as a template for his team to develop a production-ready version. SNTUC represents over 1.4 million workers in Singapore’s tripartite labor system, where unions, employers, and government collaborate on policy.

Siow began testing Microsoft 365 Copilot with 50 staff in January 2025 before rolling it out to all 1,300 employees a year later. He reports that the AI assistant has improved its ability to integrate work across Outlook and PowerPoint, reducing manual copy-pasting. Siow uses Copilot to research issues, draft documents, and prepare for meetings by summarizing agendas and gathering discussion points. His team employs the tool to translate technical solutions for non-technical audiences and create visuals for presentations.

Siow acknowledges that Copilot’s output requires human review, stating it provides a strong first draft but is not flawless. His priority is ensuring staff adoption of the tool, addressing concerns that it might go unused. With Microsoft’s support, he implemented regular training and engagement initiatives, including weekly emails with tips and best practices. About 93% of staff now use Copilot, with 74% doing so weekly, reflecting positive early feedback.

Looking ahead, Siow aims to encourage staff to experiment with AI agents to further streamline workflows. He advises organizations to focus on business value rather than technology when adopting AI, suggesting they target areas where AI can reduce friction, improve responsiveness, or enhance decision-making. His approach prioritizes practical applications over technical experimentation.