Could loyalty systems be killer context-as-a-service (CaaS) platforms?

This issue includes sponsored banners from Brevo, who also commissioned the smart loyalty report referenced throughout. The ideas in today’s editorial are related to that work, but they were not shaped or reviewed by Brevo.

Last month, I published a bespoke report for Brevo on the concept of “smart loyalty,” a more expansive view of creative engagement and reward mechanisms for loyalty programs than classic earn-and-burn discounting. The loyalty design space — social engagement, digital wallets, partner ecosystems, experiential rewards — is much richer than most marketers have explored.

This week it bumped into something else I’ve been writing about.

I’ve been hypothesizing that the next iteration of great SaaS platforms could evolve toward context-as-a-service (CaaS) — trusted, authoritative sources of domain-specific context that all the other AI agents and agentic apps in your stack could rely on.

But not all context is equal. The signals that actually matter aren’t just transactional — they’re behavioral, preferential, and intentional. What did this customer do? What do they value? Where are they heading, before they’ve consciously decided?

That raises an interesting possibility: could a well-designed loyalty program be one of the best systems in the stack for generating that kind of context?

Most of the debate about where context “lives” in a martech stack focuses on CDPs, CRMs, and data warehouses — which system is the authoritative source, which has the freshest data, which can be queried fastest. That’s a real debate, but it might be missing something.

The more interesting question isn’t where context lives. It’s where it comes from.

CDPs are flexible and powerful, but their primary role is to aggregate and unify signals that other systems generate. A full-spectrum loyalty program does something different: it actively elicits context that wouldn’t exist otherwise. The earning mechanisms, the redemption choices, the challenges, the tier progressions — these are engineered invitations for customers to reveal what they actually care about.

Consider what a rich loyalty program actually captures. Purchase patterns across channels, of course. But also product engagement (did they actually use what they bought?), social behavior (did they post, review, refer?), event participation, device signals, and — critically — reward preference signals that reveal what a customer values, not just what they clicked. Redemption choices and other voluntary loyalty interactions are disclosed preferences, zero-party data in a very literal sense.

When a customer chooses to redeem points for early product access rather than a discount, that’s not a behavioral inference — it’s a declared preference. When they complete a fitness challenge to earn points, connect their Strava account, or donate their rewards to a charity, they’re telling you something explicit about who they are and what they value.

A CDP can store that signal. But the loyalty system is what created the opportunity to generate it.

That’s a meaningful architectural distinction. A loyalty program sits upstream of the data layer by generating context that other systems simply collect. But it also sits downstream, applying that context through rewards, recognition, and next-best actions. That gives it an unusual role in the stack: not just sensing customer intent, but helping shape the loop between what a customer reveals and what happens next.

But that dual role is only as valuable as the speed at which signals move through it.

One of the underappreciated properties of context is its time-sensitivity. A signal at the moment of generation carries its maximum value. The same signal 12 hours later is history. Two days later, it may be archaeology. (Paging Indiana Jones.)

Latency determines whether a system is actually providing context or shipping stale assumptions downstream.

Consider the moment a customer completes a fitness challenge, posts a review, or makes a redemption choice that reveals something about their values. That moment is a peak signal: explicit, fresh, and directly actionable. But the value of that signal decays quickly. A system that propagates it in real time can close the loop while the customer is still engaged. A system running on batch syncs acts on context that’s already stale.

The distinction matters more as AI decisioning becomes more central to the customer experience. An AI operating on yesterday’s context has already lost the thread of what the customer is doing right now. The richness of the data matters, but so does its freshness.

A loyalty system that generates rich signals but propagates them slowly knows something valuable about this customer. It just can’t tell anyone in time to matter.

Everything discussed above — the signal richness, the context generation, the flywheel of engagement — applies just as powerfully in B2B. In some ways, it matters even more, because B2B relationships are longer, higher-value, and more complex. The cost of getting them wrong is enormous.

And yet explicit loyalty programs are rare in B2B. Not because the underlying activities don’t exist. Most mature B2B companies already invest in certification programs, user communities, executive advisory boards, early adopter programs, and customer references. The signals are there too: product usage, adoption depth, engagement patterns, and renewal trajectories.

What’s often missing is a system that connects them.

Customer success platforms do a remarkable job monitoring account health and flagging risk. But the context they generate tends to stay within the CS org — informing the CSM’s next call rather than flowing to marketing, product, or AI decisioning. The signals pool rather than propagate.

That creates a significant platform opportunity. A customer who has deepened product adoption, completed certifications, contributed to the user community, and attended an executive briefing is telling you something composite and meaningful — not through any single signal, but through the pattern across all of them.

Giving AI agents access to that accumulated context would be a powerful capability. Not just “this customer’s usage dipped,” but “this customer is deeply embedded, highly engaged, and approaching a natural expansion moment.” The difference between those two inputs is enormous.

That’s the context-as-a-service opportunity in B2B — synthesizing signals that currently live in separate systems into a coherent picture that can inform every agent and application that touches the customer relationship. The signals already exist. The architecture to consolidate and reason across them is what’s mostly missing.

The loyalty-shaped hole in B2B isn’t about points or tiers. It’s about the absence of a system designed to treat the full arc of customer engagement as a source of rich, structured, continuously updated context — and to orchestrate AI and human responses accordingly.

That’s what the best B2C loyalty systems do. B2B is still largely waiting for its equivalent.

Thank you for being a loyal reader,

Scott