AI Loyalty Programs in Singapore: Lessons from Metro

Most loyalty programmes don’t fail because the rewards are bad. They fail because they treat everyone the same.

Metro Singapore—one of the country’s most recognisable department store names—just made a very modern bet: that the next era of customer loyalty in Singapore won’t be driven by louder discounts, but by better decisions. Specifically, decisions powered by customer data, real-time analytics, and AI.

This post is part of our “AI dalam Peruncitan dan E-Dagang” series, where we look at practical AI adoption in Singapore retail and eCommerce—from personalised recommendations to demand forecasting and behavioural analytics. Metro’s loyalty refresh is a clean example of what’s changing in 2026: loyalty programmes are shifting from “points for spend” to personalised, omnichannel relationships.

Metro’s loyalty relaunch shows where retail loyalty is heading

Metro’s loyalty refresh (“Treasured by Metro”) is a signal of a bigger trend: retailers are rebuilding loyalty as a data product, not a marketing side project.

According to the source article, Metro partnered with Eagle Eye (a SaaS and AI technology provider) to support a brand and loyalty refresh across its two physical stores, eCommerce platform, and marketplace presence—plus operations beyond Singapore.

What’s notable isn’t just that there are more perks or another tier. It’s the operating model behind it:

• Customer data unified across online and offline
• Real-time decisions about who gets what offer, when, and through which channel
• Promotional earn campaigns that can flex around seasonal peaks
• A roadmap toward behavioural rewards, not just transactional ones

That combination matters because it fixes the most common loyalty problem I see in retail: a programme that looks good in a deck, but can’t execute consistently across POS, eCommerce, CRM, and marketing tools.

The real problem: discount loyalty trains customers to wait

If your loyalty programme mainly works by giving discounts, you’re not building loyalty—you’re building a habit of waiting.

Metro’s leadership called out something many retailers avoid saying out loud: department stores (and plenty of specialty retailers too) have been pulled into a loop where “loyalty” becomes synonymous with price cuts. The downside is predictable:

• Customers delay purchases until a promo drops
• Margin erodes while acquisition costs rise
• “Members” don’t feel recognised—they feel targeted

Here’s the stance I’ll take: a loyalty programme that can’t do personalisation will always drift toward blanket promotions. It’s the easiest thing to run operationally, especially when teams are stretched.

AI doesn’t magically create emotional connection. But it does reduce the effort required to do the basics properly at scale: segmentation, timing, channel choice, and relevance.

What Metro is actually implementing (and why it works)

Metro’s refreshed programme keeps a familiar structure—tiers and points—but adds capabilities that make it adaptable in the real world.

Differentiated and promotional earn (useful in Singapore’s retail calendar)

The big change isn’t “earn points.” It’s earn points differently depending on context.

The ability to run promotional earning windows means Metro can shape behaviour during periods that matter in Singapore:

• Chinese New Year gifting season
• Great Singapore Sale-style promo periods
• 9.9 / 10.10 / 11.11 / 12.12 eCommerce spikes
• Year-end travel season (when footfall patterns shift)

Instead of a blunt “20% off storewide,” a retailer can push targeted earn boosts:

• Double points on a category with excess inventory
• Bonus points for shopping across two departments (basket expansion)
• Earn accelerators for lapsed customers to reactivate

This is exactly the kind of AI in retail that quietly improves profitability: you influence behaviour without racing to the lowest price.

Omnichannel access: one loyalty identity across store + online

Answer first: omnichannel loyalty works when a customer has one identity and one set of benefits everywhere.

Metro’s updated loyalty experience emphasises consistency in both in-store and online touchpoints. This matters because Singapore shoppers don’t move in a straight line. They often:

• browse online, buy in-store (or vice versa)
• compare marketplace prices before committing
• expect service teams to “know them” across channels

If your loyalty programme breaks at any point (can’t redeem online, can’t see points at the cashier, can’t apply a perk after checkout), customers stop trusting it.

From an AI adoption perspective, omnichannel isn’t just “nice UX.” It’s what makes your data useful. Without cross-channel linking, your “personalisation” is working off partial truth.

Behavioural rewards: the underrated lever

Metro is exploring behavioural earn—rewarding actions beyond spend (for example, social sharing after attending in-store events).

This is where modern loyalty can become a growth engine, not a cost centre. Because the most valuable behaviours aren’t always purchases:

• attending a product demo or styling session
• writing a review after a purchase
• referring a friend who actually converts
• opting into WhatsApp/SMS updates (with consent)
• completing a profile (size, preferences) to improve recommendations

If you’re a retailer in Singapore, behavioural rewards also help with a very local challenge: high competition and low switching friction. Shoppers have many alternatives within a short MRT ride—or a single click.

Tiering that feels aspirational (but still simple)

Metro’s programme includes four tiers with increasing entitlements like birthday discounts, higher earn rates, free parking, and invite-only events.

Tiering works when:

1. The path to the next tier is understandable
2. The benefits are felt quickly (not only after months)
3. The top tier has at least one benefit that isn’t easily copied by competitors

Invite-only events are a strong move here. Experiences are harder to replicate than “$10 off.”

The AI piece: why “real-time decisioning” changes everything

The article describes Eagle Eye’s AIR platform as connecting customer data from Metro’s offline and online systems into a single, secure loyalty engine that analyses behaviour—what customers buy, how often they shop, and how they engage—then decides which reward/message to deliver, when, and via which channel.

Let’s translate that into plain operational impact.

Real-time decisioning beats calendar-based campaigns

Most retailers still run loyalty marketing like this:

• plan campaigns weeks ahead
• blast a segment
• review results after the fact

Real-time decisioning flips it:

• observe behaviour (browse, cart, purchase, lapse)
• choose an offer based on rules + models
• deliver it immediately (app, email, POS receipt, SMS)

That’s not just faster. It’s more honest. You’re responding to what the customer is doing, not what your campaign calendar says.

Personalisation at scale is mostly a data plumbing problem

AI in retail marketing gets overhyped, but one thing is true: personalisation fails more often due to messy data than weak models.

To make AI-driven loyalty programmes work, you need:

• a consistent customer identifier across POS and eCommerce
• clean transaction feeds (SKU, category, margin if possible)
• consent-aware communication preferences
• redemption data (what actually got used)
• latency that’s low enough to act on

Once that foundation is in place, “AI” becomes practical: propensity scoring, next-best-offer, churn risk, and customer lifetime value (CLV) become usable tools instead of dashboards nobody trusts.

Practical blueprint: how Singapore retailers can copy the good parts

You don’t need to be a 68-year-old department store to take the lesson. The pattern applies to specialty retail, grocers, beauty brands, and marketplaces.

Step 1: Decide what loyalty is supposed to change

Answer first: a loyalty programme should change customer behaviour in a measurable way.

Pick 2–3 outcomes and commit:

• increase repeat rate (e.g., second purchase within 60 days)
• grow basket size (cross-category shopping)
• reduce churn (reactivate lapsed members)
• shift sales from discount-driven to full-price

If you try to “increase engagement” as a vague goal, you’ll default back to discounts.

Step 2: Build your earn logic around moments, not just tiers

Tiers are fine, but modern loyalty needs moment-based triggers, such as:

• first purchase after signup
• first in-store visit in 90 days
• abandoned cart within 2 hours
• birthday month
• category exploration (trying a new department)

Then decide what you’ll offer that isn’t purely price:

• early access to launches
• services (alterations, styling consults)
• partner perks (parking, F&B, experiences)
• donation matching or community rewards (works well for certain brands)

Step 3: Start with “rules + reporting,” then add AI

A mistake I’ve seen: teams buy an “AI loyalty” product before they can run clean promotions.

A safer sequence:

1. Centralise loyalty data and unify identities
2. Launch clear promotional earn rules (and measure uplift)
3. Add basic segmentation (RFM: recency, frequency, monetary)
4. Introduce AI models for churn risk, CLV, and next-best-offer
5. Expand into behavioural rewards and partner ecosystems

This keeps change manageable and reduces internal resistance.

Step 4: Measure what matters (4 metrics that don’t lie)

If you’re rebuilding loyalty, track these consistently:

• Incremental revenue: sales you wouldn’t have gotten without the offer (requires control groups)
• Redemption rate: whether offers are actually relevant
• Repeat purchase rate: especially for new members
• Customer lifetime value (CLV): the north star for long-term loyalty

“Email open rate” is not a loyalty metric. It’s a communications metric.

Where AI loyalty in Singapore is going next

Metro’s direction—real-time personalisation, omnichannel identity, behavioural rewards—maps neatly to where Singapore retail is headed in 2026.

Expect more of these moves:

• WhatsApp-first loyalty communications (with careful consent management)
• Partner networks (earn in one place, redeem in another) to increase perceived value
• Retail media integration (loyalty insights informing on-site and off-site ads)
• Privacy-aware personalisation as PDPA expectations rise and customers get more selective

The biggest shift is mindset: loyalty programmes are becoming customer decision systems. Points and tiers are just the interface.

Metro’s relaunch is a timely reminder for anyone in AI dalam Peruncitan dan E-Dagang: AI is most valuable when it’s applied to a high-frequency business loop—like customer engagement—where small improvements compound week after week.

If you’re looking at your own loyalty programme and thinking “we’ve already got points,” the better question is: can you decide, in real time, what a specific customer should get next—and prove it improves CLV without destroying margin? That’s the bar now.