Why AI Agents Are Breaking Traditional Loyalty Economics
AI agents are eliminating loyalty program 'breakage' by instantly verifying offers and optimizing discounts in real time. Retailers must rebuild infrastructure for sub-250ms response times or lose customer value.
Two retail milestones in early 2026 are forcing a reckoning for loyalty programs across Australia and New Zealand. In January, Woolworths integrated Google Cloud's Gemini Enterprise into its Olive chatbot, transforming it into a full AI shopping agent. Two months later, Canadian grocery chain Loblaw embedded its PC Express service directly into ChatGPT.
Together, these moves signal that AI agents are now actively mediating purchase decisions at scale.
Woolworths and Loblaw Signal a Structural Shift
Woolworths' upgraded Olive agent supports text, voice, and image inputs. Shoppers can photograph a handwritten recipe and receive a fully built basket in seconds. The agent applies loyalty discounts in real time, checking tier status, point balances, and offer eligibility during the same interaction.
Loblaw's ChatGPT integration extends this model to third-party AI platforms. When a shopper asks ChatGPT to fill a grocery order, PC Express fulfills it. If loyalty verification cannot happen during that API call, the agent optimizes for price alone.
Adobe data shows e-commerce traffic from generative AI platforms surged 693% during November and December 2025 compared to the prior year. According to PayPal Australia research, 48% of Australians have already used AI assistants for product searches, with 78% expecting these tools to become routine.
Why Existing Loyalty Programs Cannot Keep Up
Traditional loyalty programs were built around human behavior. Customers forgot to scan cards. They missed offers. They rarely checked point balances. That forgetfulness, known in the industry as "breakage," quietly protected program economics.

AI agents eliminate all of that. They never forget. They check every balance instantly. They calculate the best available value across every competing program before a purchase is confirmed.
The technical gap is severe. Most loyalty platforms use batch processing with weekly or overnight data refreshes. AI agents require sub-250 millisecond response times for real-time offer verification. Infrastructure must also handle five times normal peak traffic simultaneously.
As one industry observer noted, "In agentic commerce, a 10-second delay effectively equals a 10-day delay. The agent has already moved on."
What Woolworths' Olive Agent Reveals in Practice
A specific mechanic documented in Woolworths' Olive integration illustrates the stakes. When a customer's cart is US$5 short of a "spend US$50, save US$5" threshold, Olive identifies the gap in milliseconds and suggests a strategic add-on. Basket size increases. The offer redeems. No human initiated any of it.
This only works because Woolworths' loyalty infrastructure can respond during the agent's API call. Retailers whose platforms cannot meet that response window are invisible to the purchase decision.
Google's Gemini Enterprise platform has also been piloted by Kroger and Lowe's in the United States, confirming this is not a single-market experiment. The agentic loyalty model is being validated across multiple retail verticals simultaneously.
Looking for World-Class PR & Comms in APAC?
Tailored service packages for select brands and agencies.
The Infrastructure and Design Overhaul Now Required
For ANZ loyalty operators, the response window is compressing. Loyalty programs must expose points as discount equivalents directly in product data feeds. Tier benefits such as free shipping must appear in API pricing, not inside member portals that require a separate login.
Deloitte and McKinsey both note that the power dynamic in agentic commerce is shifting toward suppliers investing directly in agent platforms for visibility. Retailers not engaged in equivalent pilots face a compounding disadvantage as agent algorithms increasingly prioritize loyalty data from programs already integrated into their workflows.
Google's roadmap also includes "Agentic Consent," a framework where customers authorize agents to access member-only pricing automatically. Any platform lag during identity confirmation causes agents to default to faster-responding competitors.
ANZ retailers operating on legacy batch systems face a clear outcome: agent-driven traffic will route to competitors whose loyalty infrastructure responds first.
Want to reach thousands of marketing and comms professionals across Asia?
Get your brand in front of industry decision-makers.
Partner with Mission Media →
