Consumers use world-class AI every day. Google's AI Mode passed a billion monthly users within a year of launch, and hundreds of millions more use ChatGPT, Gemini and Copilot. These tools have reset what "good" feels like. When a customer moves from that experience to a typical brand website, the drop in quality is hard to miss. That's the expectation gap.
Two shifts are driving this at the same time, and they compound each other.
- Shift one: Fewer visitors are reaching brand websites at all. AI assistants increasingly answer the question on the results page itself. In the first four months of 2026, 68% of Google searches in the US ended without a click, up from around 60% two years earlier. Google's own numbers show the effect: search revenue grew 19% in the quarter even as fewer of those searches sent people on to the open web.
- Shift two: The visitors who do arrive expect more and forgive less. Even before generative AI, 69% of shoppers went straight to a site's search bar, and 80% left when the experience did not meet their expectations. Raise that bar to the level of a daily AI habit, and the visitor who still reaches a brand's site is both more valuable and less patient than before.
Does this mean brand websites are disappearing?
No, but their role is changing.
AI search will absorb more of the functional work websites used to do — answering basic questions, comparing options, narrowing choice — acting almost like a concierge.
When someone does choose to visit a brand's site, that visit likely will carry more intent. Customers may arrive after researching elsewhere, with clearer needs and sharper questions about whether the brand is right for them. Those visitors will be looking for reassurance, perspective and confidence.
That raises the standard for what a brand website must provide. A website is no longer the default destination for discovery. It must offer a concierge-style experience: understanding each visitor's context, anticipating what they need and doing the heavy lifting to help them get there.
How to assess your own brand's vulnerability to this shift
You are exposed to the expectation gap if any of the following is true:
- Your site still returns keyword-matched links instead of direct answers. Customers have learned, from the assistants they use daily, that they can ask a question and get an answer. A results list that makes them do the matching themselves now reads as dated, not neutral.
- Your analytics show declining organic sessions with stable or improving rankings. That combination is the fingerprint of zero-click behavior, not a ranking problem.
- A meaningful share of your category's research happens conversationally. In one 2026 study of active AI users, 37% of consumers said they now begin their search in an assistant rather than a search engine, 40% said they were tired of clicking through links and 47% said AI-generated answers already influence which brands they trust first.
It is less likely to apply, at least urgently, if your business runs on high-frequency repeat visits from an app or logged-in relationship rather than open search discovery. The expectation gap is widest wherever a customer's first touch is a search or a question.
Does closing the gap mean you need to overhaul your data first?
No, and this is the assumption that stalls most teams.
The concern that AI features will produce poor or inaccurate answers without extensive data foundations is valid in some contexts, but it gets applied more broadly than it should be.
For public website content, much of the groundwork already exists. If a site is already indexed by a search engine, that indexing is a foundation, not a blocker, for adding semantic and generative search on top of it.
Robust data foundations matter most when an experience needs to give definitive answers from private or sensitive information, such as account data, pricing exceptions or regulated content.
For improving discovery on content that is already public, most teams are closer to a working version than they assume.
Where should a brand actually start?
This is the question that matters most, and it does not have a single right answer. Closing the expectation gap is a series of steps, not one leap, and the right stopping point depends on your brand's goals and risk appetite.
Step 1 is a reasonable place for most brands to begin. It's what Valtech and Google helped Dunelm achieve. That experience is available today and already represents a significant improvement on conventional site search.
How urgent is it for brands to act?
Make this a near-term priority. Two developments are unfolding right now that create urgency for brands everywhere:
- How these experiences are packaged. A chat window is not the only option, and it is not right for every brand. Much of the current innovation is in what we would call adaptive UI: a response generated dynamically for the moment, personalized, and shaped by the brand's own design and tone, rather than a list of links or a wall of text. The underlying AI capability is well established. The design work now happening is in the experience layer around it.
- Who is doing the visiting. A growing share of traffic will not be people at all, but agents acting on a person's behalf. Agents are unforgiving of friction. If one cannot navigate a site's flow, it moves on, often without leaving a clear trace in the brand's analytics. Being readable to a machine is becoming as important as being appealing to a person.
So, what should a brand do this quarter?
Start with an honest comparison. Put your own experience next to the AI tools your customers use daily, and look at the distance between them.
From there, it's more productive to choose a step 1–4 to start from than wait for perfect data conditions. Step 1 is a reasonable place to begin for most brands, and each step built well makes the next one easier.
If you need help identifying where to start, book a maturity assessment with our team.
