The intelligence sitting between data and activation is improving fast. But a decisioning engine is only as good as the signals feeding it, and for most brands, the most important signals are the ones that never get captured. This piece looks at the real-world data gap, what it costs as AI takes on more of the personalisation layer, and a channel that's already in the consumer's hand.
One theme kept surfacing at MartechDay 2026: the decisioning and orchestration layer, the intelligence sitting between data and activation, is finally getting the attention it deserves. The ability to decide who gets what experience, at what time, through which channel, is maturing fast. AI is making real-time personalisation at scale actually achievable. The tooling is catching up to the ambition.
The problem sitting underneath all of it? The middle layer is only as good as the data feeding it.
Most customer data stacks are well-optimised for digital behaviour. Email opens, web sessions, app activity, ad interactions. These signals get captured, unified, and fed into decisioning engines with increasing sophistication. The infrastructure for understanding what happens on a screen has become genuinely impressive.
What's largely missing is everything else.
The majority of consumer decisions happen away from screens. Picking up a product in-store. Showing up at a brand event. Walking past a venue. For most brands, those moments pass without generating a single usable signal, even though they're often where purchase intent is highest. A customer spending two minutes in a category aisle comparing labels is doing something commercially meaningful. Most brands have no record of it happening.
The result is a personalisation engine working from an incomplete picture. It knows someone clicked an email three weeks ago. It doesn't know they were standing in a competitor's aisle yesterday, or that they've attended three of your sponsored events this year.
For FMCG brands specifically, this gap isn't a minor inefficiency. Brands selling through retail have operated for years on aggregated, delayed data borrowed from distributors. They've never owned a direct line to the individual consumer. The person buying their product has been a demographic estimate, not a known person with a live profile. When a decisioning engine tries to personalise for that consumer, it's working from something closer to an assumption than a record.
That's not a decisioning problem. It's a data capture problem. And a better orchestration engine alone won't fix it.
Already in their pocket
Apple Wallet and Google Wallet are on virtually every smartphone. Pre-installed, trusted, used daily. Most people open their wallet multiple times a week without really thinking about it. It's not an app competing for a download. It's already there.
For brands, that's a direct channel sitting in the consumer's hand at every real-world moment that matters. The question is whether they're using it.
A wallet pass is not a static card. It's a live connection. Content updates in real time. Location triggers fire geo-targeted offers the moment a customer steps within range of a store, venue, or event. Notifications land directly on the lock screen. Every tap, redemption, or response flows back as an individual-level signal tied to a real person at a real moment.
The data that comes back is specific: this person, this location, this time, this action. For a brand that has historically received a monthly distributor report summarising category-level sales volumes, that's a fundamentally different kind of asset.
The mechanics are straightforward. A consumer encounters a brand in the real world, and an incentive drives them to add a wallet pass. From that point, the brand has a direct channel and a profile that builds with every subsequent interaction. When they walk near a stockist, a geo-targeted offer fires to their lock screen. When they redeem it, the signal goes straight back into the stack. The pass updates: points accumulated, next reward available, a new challenge activated. The relationship continues past the initial interaction. The profile deepens.
Event and venue operators have a similar opportunity that most aren't yet using. An attendee adds their ticket to their wallet and, as they move through the event, location-triggered activations fire in sequence: a hospitality offer at the bar, a post-show reward on the way out, a challenge running across a multi-day festival. That data doesn't sit isolated in a ticketing platform. It builds a profile that gets sharper with every touchpoint and persists long after the event.
Loyalty programmes follow the same logic. The signal is immediate, individual, and owned by the brand. The member stays in a live relationship with the programme rather than a static card they eventually stop carrying, or a phone number they recite at the counter while hoping their points are still there.
AI doesn't close the gap. It amplifies it.
The reason this matters more now is where the AI conversation is heading. The agentic layer, intelligent systems making real-time decisions about how to engage each individual, needs context to work well. It needs a continuously updated picture of what a customer has done, where they've been, and how they've responded. An agent working from a thin or stale profile will optimise confidently in the wrong direction. It will reach people with the wrong message at the wrong moment, with complete conviction. More processing power applied to incomplete data doesn't close the gap. It widens it.
Brands building that foundation now, capturing real-world signals alongside digital ones, will have something genuinely useful to hand those systems. Brands that haven't will find that smarter tooling just amplifies what's already missing.
The wallet is one of the most direct ways to close it. The infrastructure is already in the consumer's hand. The question is whether brands start treating it as a data channel rather than a card.
