A category that claims to be for everyone is a category no one believes. So this paper does something vendor literature rarely does: it draws the boundary honestly, names who falls outside it, and tells some readers — candidly — that Enterprise Customer Memory is not for them.
The boundary is not industry, size, or budget. It is a kind of business.
Recognition businesses
Some businesses are built on transactions: the sale is the relationship, and the next sale starts from zero by design. Others are built on recognition: the customer returns, the business is expected to know them, and being known is part of what the customer is paying for. A luxury hotel. A restaurant with regulars. A private bank. A clinic. A dealership's service desk. A members' club.
In recognition businesses, memory is not a nice-to-have layered onto the product. Memory is the product, or close to its heart — which is why these businesses have always run informal memory systems at great cost: the maître d's book, the concierge's index cards, the banker who "knows the family." Enterprise Customer Memory is for recognition businesses, and only for them.
Four questions determine whether a business is one. All four must pass.
1. Does the enterprise meet the customer directly, and repeatedly? Not through a retailer, not anonymously, and not once. If the relationship is mediated or single-shot, there is nothing to remember and no one to remember it for.
2. Is there an identity spine — or a credible staircase to one? Memory attaches to identity. An email, an account, a sign-in worth the customer's while. A business whose customers are cash transactions at arm's length has records it can never resolve into people.
3. Does knowing the customer change the very next interaction? Memory must have a moment to matter in. If nothing the business would do differs for a known customer versus a stranger, understanding is trivia.
4. Is there an owned surface at the moment of interaction? A website, an application, a host stand, a front desk, an agent — somewhere the memory can be delivered when it counts. Understanding with nowhere to arrive is a diary.
Run the test on a consumer-goods company selling through retail: the customer buys at a chain store, anonymously, every few years; there is no account, no moment, no surface. Zero for four — and no AI initiative changes that arithmetic. Run it on a hotel: direct, repeated, identified at booking, with a dozen owned surfaces from the website to the concierge. Four for four, with room to spare. Restaurants, clinics, wealth managers, dealerships, private clubs: four for four. The category is large — and bounded.
Within hotels: the second fence
Hotels pass the recognition test as a class. But within hotels, Enterprise Customer Memory is for upscale, upper-upscale, and luxury properties — and this fence is economic, not snobbish.
The recognition premium has to be absorbed somewhere. It monetizes through rate — guests who are known become less price-sensitive — and through ancillary spend at the moment of intent: the spa, the restaurant, the golf course, the upgrade, the experience. A full-service property has a dozen surfaces where being known converts to revenue. A limited-service property has a lobby and a keycard; it can implement memory and have nowhere for memory to earn. The fence follows the money.
There is a simpler way to say all of this, and it is the test we actually use:
What if your hotel could remember every guest like Aman?
Aman does it with headcount. Enterprise Customer Memory does it with architecture.
Aman is what recognition looks like when it is the entire brand promise: guests describe being known across continents, and the industry has treated that as the unreachable summit — achievable only with fifty keys, extraordinary ratios, and staff longevity no ordinary property can sustain. That summit is a headcount achievement. The claim of this category is that the memory half of it — not the service, the knowing — is now an architecture achievement, available to a 300-key property that could never staff its way there.
If that question stirs something in how you think about your hotel, you are inside the fence. If it reads as irrelevant to your operation, you are outside it — and this category, honestly, is not for you.
Why we publish the boundary
Three reasons, stated plainly.
Because a bounded claim is a believable one — the fastest way to be doubted about who needs a thing is to claim everyone does. Because the fence protects buyers as much as the vendor: a deployment where the economics cannot absorb the premium fails, and a failed deployment teaches an industry the wrong lesson about a young category. And because categories are defined as much by their edges as their centers: CRM became legible when it was clear what wasn't CRM.
The market inside this fence is worldwide — every upscale-to-luxury property on the planet faces the same nine recording systems and the same returning stranger. Tens of thousands of properties. Not everyone.
That's the point.