By Marcus Puffer, Vice President & Head of Loyalty Solutions, IBS Software and Binay Warrier, Vice President, Loyalty Management Solutions, IBS Software
For decades, loyalty programs have been built around a simple premise, to influence human behaviour. Points, tiers, perks, and personalized offers are all designed to nudge people toward choosing one brand over another. But that premise is beginning to crack. As purchasing decisions increasingly shift to AI agents acting on behalf of consumers, loyalty is entering a fundamentally new ere, one where the “buyer” is no longer human.
Now, you could ask your AI travel agent to book a flight from London to LA. It knows your preferences, your memberships, and your points. The agent reaches out to Airline A and Airline B on your behalf. Airline A has the cheaper fare. Airline B’s loyalty platform signals that you are one flight away from Platinum status, and mentions a targeted incentive, such as increased discounts on mile conversions, to influence the agent to look at overall value, not just price. The agent recommends airline B, and if you have already delegated purchasing authority, it books without asking.
This is agentic commerce, a model in which AI agents are given real decision-making authority on behalf of consumers. These agents are actively evaluating options and executing purchases based on a user’s preferences, history, and goals, and agentic commerce is arriving faster than the loyalty industry realises.
What does loyalty mean when the buyer is an algorithm?
Traditional loyalty programs work, in part, because humans are emotional. irrational. Travellers choose familiar airlines even when cheaper options exist, driven by habit, status, or emotional attachment. AI agents are not sentimental. They optimise, and prioritise tangible benefits, leaving out the emotional component, and make decisions accordingly.
But that doesn’t mean loyalty becomes irrelevant, but it must evolve. Loyalty programs are not themselves an emotional experience, that comes from product, service, and experience. A loyalty program is the data infrastructure that enables brands to design those experiences consistently, and the memory layer that reinforces why a customer once chose that brand.. What changes is how that value is communicated. There is a fundamental gap between what makes a human choose a brand (status aspiration, habit, emotional attachment) and what makes an AI agent choose a brand (quantified value, contextual advantage, speed of response). This is the persuasion gap. Bridging it means shifting from emotional persuasion to computational persuasion, delivering real-time, personalised, contextual value arguments that are as compelling to an algorithm as brand loyalty is to a human.
In an agentic world, that influence must be translated into something machines can understand and act upon. An AI agent embedded in ChatGPT or Copilot already carries a deep understanding of its user’s preferences and behavioural patterns. The question is whether loyalty platforms can plug into that understanding and bridge that persuasion gap, or remain invisible to it.
The machine-readable gap nobody is talking about
Today’s loyalty technology was built to communicate with humans, through apps, and emails on a desktop or mobile screen. Naturally, most programs were not built to expose their value in a structured, machine-readable way.
Yet in the age of AI assisted commerce, while an AI agent can easily compare prices and schedules, it cannot understand upgrade probabilities, or personalised incentives unless that data is deliberately brought to the surface for it to read.
Airlines are therefore risking loyalty becoming invisible in agent-driven decisions.
The technical foundation that closes this gap already has a name, Model Context Protocol (MCP). It allows external agents controlled, structured access to a platform’s data and logic, enabling a booking agent to query a loyalty program and receive back a member-specific value proposition in real time. Emerging agent-to-agent protocols will take this further, allowing specialised loyalty agents to actively participate in the purchase decision and communicate proactively with booking agents on a member’s behalf, highlighting proximity to a status tier or a tailored redemption offer at the exact moment it matters most.
In effect, loyalty programs would no longer just influence humans. They would negotiate with machines.
Short-term not long-term shift
This is not a distant fantastical shift. Agentic commerce is already happening, in business transactions, automated procurement systems have long used algorithms to optimise purchasing decisions. In consumer contexts, AI assistants are rapidly gaining capability, from planning trips to recommending specific services. The missing piece for airline loyalty programs is execution, and that gap is closing fast.
That urgency has a direct competitive implication. If Airline A builds MCP-ready loyalty infrastructure and Airline B does not, Airline A will be highly visible in agent-mediated decisions, Airline B will be ignored. Multiply that across millions of agent-assisted bookings and the gap compounds in revenue, and in long-term relevance.
When Google become a disruptor between brands and consumers, allowing airlines to drive customers into their direct booking channels, airlines that moved slowly spent years trying to reclaim direct relationships at significant distribution cost.
Agentic AI enabled through Agent-to-Agent protocols is the next intermediary, except this time, this one doesn’t just inform, it acts. Airlines that wait risk the same outcome where the direct member relationship will once again be mediated by a third party. Agentic AI exposure will become a new equivalent to SEO on steroids, determining not just visibility, but whether a brand is considered at all in the buying equation.
Preparing for an agentic shopping reality
The implications are both technological and strategic. On the technology side, organisations need platforms that support real-time data exposure, API-driven interactions, and intelligent agent layers. On the strategy side, the core question is no longer just about influencing customer behaviour, but also about ensuring that AI agents are making decisions that consider the rational and emotional decision factors for a customer. These factors are significantly influenced by a loyalty programs offering. The challenge here is for loyalty programs to build a strong proposition that is reflected, understood, and chosen by a customer’s AI purchasing agent. The winners will bridge the persuasion gap on both sides, delivering computational persuasion compelling enough for the machine and emotional experiences that the human still feels brand loyalty.
If a customer’s AI agent reached out to your loyalty program today, what would it find? In the age of agentic commerce, loyalty isn’t disappearing, it’s being re-written in an entirely different language.