Loyalty programs were built on a simple promise: know your customer, reward their commitment, make them feel seen. Now AI agents are doing the shopping, which raises a question most loyalty leaders haven't fully answered: if the machine handles the transaction, who handles the relationship?
The transaction has left the building
According to Morgan Stanley research by 2030 nearly half of all online shoppers will use AI agents to make purchases on their behalf, adding US$115 billion to US ecommerce alone. Amazon's Rufus powered a 100 percent surge in purchase sessions during Black Friday compared to trailing averages, according to the same research. Walmart and Target are live inside ChatGPT. Macy's "Ask Macy's" AI stylist drove customers to spend around 400 percent more than non-users, according to Bloomberg. Google's Universal Commerce Protocol (UCP) will make loyalty benefits machine-readable by AI agents, brands not structured for agent-readability won't just miss the sale, they'll be invisible when the decision is made.
If an AI agent is recognizing your membership, applying your points and completing checkout, what role does loyalty actually play? Is the future of loyalty just a technical infrastructure problem?
The journey your members are on isn't the one you mapped
Most loyalty programs were designed around journey maps built for human touchpoints. In almost every audit I've run, the map and the reality are different documents. The map shows a clean earn-and-redeem arc.
The reality includes AI-driven discovery, off-platform purchasing, automated service loops and algorithmic triggers that nobody planned for and nobody is managing. That gap is where loyalty leaks.
But consumer behaviour hasn't shifted as fast as the technology.
According to the Australian Loyalty Association's 2025 Loyalty Report, 62 percent of customers still expect brands they shop with regularly to know their preferences, and 72 percent expect better service if they're part of a loyalty program.
The expectation of being seen hasn't gone anywhere. What's changed is who, or what, is responsible for delivering it.
The moments that actually build loyalty
Grill'd rewards members when their AFL team wins, Go Dees!, a tribal moment no algorithm triggers. Mecca's Beauty Loop gives top-tier members experiences that feel curated, not generated. These aren't loyalty mechanics. They're emotional decisions.
Working with Medibank, I saw this directly. Journey analysis showed the highest satisfaction scores consistently occurred at personalized, memorable moments, almost always with a human touch. The moments that mattered most weren't the frictionless ones. They were the ones where the member felt genuinely seen and understood.
That insight led me to redesign the welcome experience: every new member received a personalized welcome video. The technology made it scalable. The human thinking made it land.
Artificial intelligence (AI) and human moments serve different jobs. AI owns recognition, hyper personalization, redemption, replenishment and agent-led checkout. Humans own recovery after failure, milestone moments and the experience that makes someone tell a friend.
According to Qualtrics' 2026 Consumer Experience Trends Report, nearly one in five consumers say AI customer support delivers no benefit at all, the worst rating across all AI tasks measured. Frictionless is not the same as meaningful.
Identifying where to automate or humanize
The reason most brands struggle with this isn't ambition. It's that they don't have a structured way to decide which moments should be AI-powered and which must stay human. That's the problem I aimed to solve when I built the ACTIVATE Framework™
A - Audit the real journey
Map the journey customers are experiencing including AI-influenced touchpoints, automated loops and algorithmic triggers. If you don't map them, you can't manage them.
C - Clarify moments of disproportionate importance
Not every touchpoint deserves equal investment. Data should tell you where emotion peaks, trust forms, and churn decisions are made.
T - Triage AI opportunities
Before deploying AI in the member journey, ask: does this reduce friction without removing meaning? Deploy with purpose, not novelty.
I - Integrate data, CX and marketing
AI fails in silos. Loyalty programs disconnected from service platforms and CRM cannot deliver the contextual intelligence that makes personalization feel real.
V - Validate through commercial metrics
Move beyond engagement metrics. Track retention lift, redemption rates among high-value members and lifetime value. If AI investments can't connect to these numbers, they're experiments, not strategy.
A - Assign accountability
Without clear ownership, AI initiatives drift. Someone must own the end-to-end member journey, not just the program mechanics or the technology, but the experience that sits across both.
T - Test in priority journeys
Don't deploy AI everywhere at once. Start where friction is highest and commercial upside is measurable, the welcome journey, the redemption experience, the lapsed member flow. Prove the outcome, then scale.
E - Embed the human overlay
This is the step most programs skip. Automation should remove friction, not empathy. The Medibank welcome video worked because a human decided it mattered – the technology just made it possible at scale. Loyalty was never really about points. It was about the feeling a brand could create. You can't automate that, but you can design for it.
The loyalty question worth asking
The future of loyalty is not human or AI. It's knowing the difference between a moment that can be automated and one that must be felt.
If AI becomes the operating system of loyalty, human moments become the differentiator. The brands that understand that and design for it deliberately, are the ones worth being loyal to.
Quick links
- The human-machine translator: CX's most important new role
- Loyalty program fraud: How points can become a shadow currency
- How the WNBA is redefining fan engagement through community-led experiences