Your AI Agent Is Your New Loyalty Card

Shawn Tan

8 hours ago

3 min read

By Shawn Tan, Interactive Rewards Asia

Published 15th September 2026

The Problem

A member of a Southeast Asian telco loyalty program calls in about a billing error. The chatbot asks for an account number, then asks again two messages later, then hands the member to a queue with a forty-minute wait. The same member has enough points sitting in the account for a free device upgrade, but nobody mentions it. This is what a loyalty program looks like when the support function and the rewards function are built by two teams that do not talk to each other.

The member does not remember the points balance. The member remembers the wait.

That gap, not the rewards catalogue, is now the biggest loyalty risk on the table.

Why This Matters Now

Adoption of artificial intelligence (AI) agents in customer service has accelerated sharply over the past year. Salesforce's State of Service research, based on a global survey of more than 3,000 customer service professionals, found that 66 percent of service organizations now deploy AI agents, up from 39 percent a year earlier. In the APAC region, this shift is landing on top of a loyalty market that is already growing fast: one widely cited market research estimate puts the region's loyalty sector at a 17 percent annual growth rate between 2021 and 2025, with double-digit growth forecast to continue through 2030. Southeast Asian and Indian markets are cited as leading this growth, driven by mobile-first consumers who now expect fast resolution as a baseline, not a bonus.

The two trends are colliding. Members increasingly judge a loyalty program by how fast a problem gets solved, not by how many points sit in the account. A loyalty program that responds slowly is, in practice, a loyalty program that is losing value in the member's eyes, regardless of what the points catalogue offers.

Where Most Teams Get It Wrong

Two budgets, no shared data
The AI chatbot lives in the contact center's budget. The rewards catalogue lives in the marketing team's budget. Neither team can see what the other is doing, so the AI agent cannot see a member's tier status, redemption history, or points balance while handling a query.

Measuring cost, not retention
Teams report ticket deflection rate and cost per contact as success metrics, while member churn following a poor support interaction keeps climbing. A member who cannot resolve a billing question in one conversation is a churn risk, regardless of how many points sit unredeemed in the account.

What To Do Instead

Connect the agent to loyalty data

Give the AI agent access to loyalty data, not just support scripts. DBS Bank rolled out an enhanced version of its generative-AI assistant, DBS Joy, to all corporate clients in November 2025. Since trials began in February 2025, the assistant has handled more than 120,000 conversations, with satisfaction scores up 23 percent during the trial period. The system works because it draws on the bank's existing client data, not a script disconnected from the account.

Surface value inside the conversation

Let the AI agent offer relevant loyalty value inside the support conversation itself, not as a separate marketing message sent later. In April 2025, Grab launched agentic AI tools for merchant and driver partners, built with OpenAI and Anthropic, that give real-time, specific guidance rather than generic responses. The same logic applies to member-facing support: an agent that can see a member is close to a reward tier can mention it in the same conversation as resolving a complaint, turning a service interaction into a retention moment.

Give CX a seat in AI design

Assign a loyalty or customer experience leader to own AI agent design, rather than leaving it solely to information technology or contact center operations. AirAsia has built AI-powered self-service tools to handle a high volume of member queries directly, an approach worth studying for any loyalty program still routing every support interaction through a generic vendor script with no connection to the rewards side of the business.

The Outcome

When support and loyalty data connect, the effect compounds. Faster resolution reduces the immediate churn risk. Relevant, in-context reward messaging lifts redemption rates and repeat engagement without a separate campaign. Programs that connect these two functions typically report gains in both customer satisfaction scores and program engagement within two to three quarters, without adding headcount to the support function.