A hotel may describe itself as romantic, peaceful, family-friendly or upscale. A tour operator may highlight its expertise, its support or the quality of its itineraries.

That is natural: every business needs to articulate what it wants to stand for. But that promise never exists in isolation. Alongside the brand's messaging is everything customers say. What they notice, what stays with them, what they find remarkable — or not. And when we look at how a business is understood in AI environments, this gap between what the brand claims and what customers say becomes particularly interesting.

Ultimately, a good rating tells us very little about why a property should be recommended. A hotel may score 4.6 out of 5 because its staff are attentive, its beach is pleasant, its rooms are comfortable or it offers excellent value for money. The rating does not tell us whether it is genuinely perceived as romantic, peaceful, family-friendly or particularly well suited to a wellness break. Yet these are precisely the attributes that may matter when a traveller makes a specific request.

The right question is therefore not simply: are customers satisfied? It becomes: do customers confirm the reasons we want to be chosen?

This is where Voice of Customer takes on another dimension. Take a hotel that wants to be recognised as a place particularly suited to couples seeking peace and quiet. Its website may be very clear, its photos may convey that impression, and its partners may echo that promise. But if customer reviews focus mainly on dining, entertainment or value for money, with very little mention of tranquillity or privacy, there is something worth examining.

That does not mean the positioning is wrong. Nor does it mean that AI will automatically penalise it. It simply means that a promise important to the business rarely appears in the accounts of people who have actually lived the experience. And before investing in more content to reinforce that promise, it is worth asking a few questions. Why do customers say so little about it? Is the experience not distinctive enough? Does it exist but fail to leave a lasting impression? Does another aspect of the hotel dominate their perception? Or do customers simply use different words to describe it?

We encountered this type of situation during a recent audit. An attribute was strongly emphasised across several sources describing the property; on paper, it should have been an advantage in certain recommendation scenarios. Yet that attribute barely appeared in the customer feedback we examined. At the same time, the property was rarely selected in several situations where this advantage could have mattered.

We cannot conclude that one directly explains the other. Generative engines draw on many signals, and their mechanisms remain too complex to attribute a result to a single cause — which is precisely why we prefer to interpret these situations through a set of recommendation factors rather than an isolated criterion. But the gap was clear enough to raise a far more useful question than “should we produce more content?”: should we communicate this promise more extensively, or should we first understand why customers say so little about it?

We find this question interesting because it forces us to stop treating reviews solely as an indicator of satisfaction or online reputation. They also become a way to test the strength of a positioning. And sometimes, contradictions are more useful than confirmations.

If a property promises peace and quiet and customers regularly mention noise, the issue is obvious. But the gaps are often less dramatic: a promise may be everywhere in the brand's messaging and virtually absent from reviews. That absence is not proof, but it is a signal.

Conversely, customers may bring to light something the business itself barely notices: a thoughtful gesture from staff, an atmosphere, a particularly smooth part of the customer journey, a very specific experience, a feature that marketing considers secondary but that customers mention spontaneously and consistently. In this case, the customer voice no longer serves only to check whether the promise holds true: it can reveal another source of differentiation. This is particularly important in tourism, because much of an experience's value is difficult to capture in a product description. The actual experience often extends beyond the intended positioning.

To explore this issue, we use a fairly simple assessment grid. For each important attribute, we look at whether it genuinely forms part of the positioning, whether it is well documented across the brand's materials, whether it emerges spontaneously in customer feedback, whether other sources confirm it, and whether it appears in the contexts in which AI recommends the business.

This analysis reveals very different situations. An attribute may be strategic, well documented and strongly confirmed by customers: that is a robust signal. It may be strategic but rarely appear in reviews, in which case we need to understand why. And sometimes, customers highlight an attribute the business had never really considered an area of differentiation.

This is where the link between GEO and Voice of Customer becomes interesting. The issue is no longer simply how to gain greater visibility in AI. It becomes: what genuine reasons are there to recommend us, and does the lived experience confirm them? This is a much more demanding question, but also a much more useful one — and it is exactly what we seek to resolve when we apply the See, Align, Build, Amplify method.

A promise becomes truly powerful when it is no longer simply asserted by the brand, but spontaneously described by those who have experienced it.