Singapore insurer Singlife says its use of Salesforce AI agents is beginning to show measurable operating results, with one customer-service agent handling more than 20 percent of email enquiries and another cutting average handling time by 30 percent.
The update builds on Singlife’s deployment of its internal AI agent Buddy, which TNGlobal covered in October 2025. According to new materials shared by Salesforce and Singlife, Buddy has also reduced training time for new customer-service hires by 50 percent, taking onboarding from four weeks to two.
Email agent handles more than one-fifth of enquiries
Singlife’s General Insurance Email Agent assesses incoming customer emails in real time, identifies the nature and urgency of each request, and routes the enquiry to the appropriate support path. Straightforward questions can receive an immediate response, while more complex or sensitive cases are escalated to human agents.
The insurer said the system now handles more than 20 percent of customer enquiries received by email. During a period of increased email volume earlier this year, Singlife said it was able to resolve most email enquiries within two days.
The operating model illustrates a common approach to customer-facing AI deployment: automate routine classification and responses while keeping human escalation available for cases that require judgment, sensitivity or more complex policy handling.
Buddy moves toward customer-facing use
Buddy was introduced last year to support Singlife customer-service executives with real-time responses to internal questions. The latest update says the agent has reduced average handling time by 30 percent and cut new-hire training time in half.
Singlife plans to move Buddy beyond its current employee-assistance role and develop it into a customer-facing digital agent. The insurer said policyholders would eventually be able to use it across digital channels to access policy information and complete routine self-service requests.
That shift will place more emphasis on how the system handles identity, permissions, escalation and the boundary between routine self-service and decisions that require an employee. It will also provide a clearer test of whether the productivity gains reported internally translate into a better customer experience when the agent is exposed directly to policyholders.
ASEAN enterprises move from pilots to operating metrics
Salesforce has been highlighting production deployments across Southeast Asia as it expands Agentforce and related AI interfaces. The Singlife figures are useful because they move the discussion from product availability toward operating measures such as handling time, training time, enquiry share and service levels.
For Singlife, the next phase will be less about whether an AI agent can answer questions and more about how reliably it can work within insurance-service controls as its role expands. The performance figures in the latest update were supplied by Salesforce and Singlife and have not been independently audited by TNGlobal.
Singlife partners Salesforce to launch AI agent for customer service executives