Singtel and Sierra Launch Singlish-Savvy AI Agents, Solving 70,000 Cases in Six Weeks

Singtel and Sierra's AI agent Shirley handles chat/voice, gets Singlish, and lets customers self-serve fixes and add-ons. Early results: 70k cases, 73% fixes, 76% roaming.

Categorized in: AI News Customer Support
Published on: Mar 10, 2026
Singtel and Sierra Launch Singlish-Savvy AI Agents, Solving 70,000 Cases in Six Weeks

Singtel and Sierra Roll Out AI Agents Customers Actually Want to Use

Singtel partnered with Sierra to upgrade customer engagement across the group, starting with a pilot in Singapore. The AI assistant, "Shirley," now handles natural, conversational chat and voice interactions so customers can resolve issues and complete transactions on their own.

"Shirley" understands local expressions-including Singlish-and more languages are on the roadmap to better serve different communities. The focus: faster resolutions, less friction, and a support team freed up for complex work.

What Changed in Singtel's Support Stack

The pilot began in late January inside Singtel Singapore's customer care. Sierra's agentic AI verifies identities, pulls relevant product and service info, and completes tasks inside virtual support channels-accurately, securely, and with human oversight built in.

Voice-based agents are next for outbound sales calls, with strict compliance and governance. The plan is to extend these capabilities across departments and offer them to enterprise clients.

Early Results Support Leaders Care About

  • Launched in under 10 weeks; 70,000+ customer cases handled in the first six weeks.
  • 73% of mobile and home troubleshooting resolved without a human officer.
  • 76% of roaming sign-up requests completed end-to-end by customers.
  • 200+ roaming add-ons purchased independently.

Containment is high, response times are shorter, and agents focus on edge cases where human judgment matters. That lifts productivity and strengthens customer trust.

Leadership Signals

Ng Tian Chong, CEO of Singtel Singapore: "Agentic AI across chat and voice is about real value-more personalized experiences, greater convenience, and faster, more responsive service... Our AI agents are designed to understand local expressions and even Singlish."

Bret Taylor, CEO and Co-Founder of Sierra: "AI can create better, more human customer experiences... Our partnership with Singtel is deepening relationships with customers while growing the business."

Why This Matters for Your Contact Center

This rollout shows what "good" looks like: local language fluency, tight policy controls, quick time to value, and a clear handoff to humans when needed. It's a practical pattern you can copy without boiling the ocean.

A Practical Playbook You Can Adapt

  • Pick 3-5 high-volume intents first (billing, roaming, plan changes, device issues).
  • Localize thoroughly: include slang, code-mixing, and common misspellings in training data.
  • Wire compliance from day one: consent capture, call recording flags, audit trails, and opt-out paths.
  • Ground responses in approved knowledge with linkbacks; block free-form claims and set strict escalation rules.
  • Design clear fallbacks: confidence thresholds, smart routing, and warm transfers with full conversation context.
  • Track what matters: containment, AHT, CSAT, FCR, deflection to self-service, and re-contact rates.
  • Upskill your team to supervise AI: prompt reviews, QA workflows, and exception handling.
  • Pilot voice after chat: start with straightforward outbound use cases (renewals, add-ons) with strict consent handling.

Risk and Governance Checklist

  • Identity verification before account actions; mask and minimize PII exposure.
  • Policy guardrails to prevent off-script answers; cite sources for sensitive info.
  • Human-in-the-loop for edge cases, complaints, or vulnerable customers.
  • Continuous QA: sample reviews, red-team prompts, and incident playbooks.
  • Clear customer signals: who they're talking to (AI vs human), what's recorded, and how to reach a person fast.

What Singtel's Move Signals for Operations

AI agents can take a meaningful share of repetitive volume if they're integrated with core systems, speak like your customers, and operate under tight controls. The payoff is fewer queues, more accurate resolutions, and agents doing higher-value work.

Learn More and Level Up

The signal is clear: start small, ship fast, measure hard, and keep humans in control. That's how you make AI work for your customers and your team.


Get Daily AI News

Your membership also unlocks:

700+ AI Courses
700+ Certifications
Personalized AI Learning Plan
6500+ AI Tools (no Ads)
Daily AI News by job industry (no Ads)