When customers ask AI instead of Google
For twenty years, being found meant appearing in a list and earning the click. Increasingly the answer arrives assembled, with two or three sources named underneath. If you are not one of them, you were not in the room at all — and you will not see it in your traffic reports, because there was never a visit to count.
This has a clutch of new acronyms attached — answer engine optimisation, generative engine optimisation. Ignore the vocabulary. The mechanical change is simple: the interface between your customer and their answer is now a summariser, and the summariser picks a very small number of sources.
What this actually changes
Traffic stops being the measure. A customer can read your opening hours, your price range and a summary of your reviews without ever loading your site. That is not lost business — they may still walk in — but it is invisible business. Judging your marketing on sessions alone will increasingly mislead you.
Being mentioned elsewhere matters more. These systems assemble answers from whatever they can read about you: your site, directories, review platforms, news, forums, industry listings. A business that exists only on its own website is thin evidence. One that appears consistently across several independent places is easy to summarise confidently.
Consistency becomes a ranking factor in practice. If your address is written three different ways across four platforms and two of your phone numbers are dead, a summariser has to pick one and may pick wrong — or skip you for a source it can reconcile.
Optimising for a machine that reads has turned out to mean the same thing as being genuinely clear: say what you do, where, for whom, in plain sentences.
What appears to earn a citation
- Answers phrased as answers. A page that states a specific question and answers it directly in the first two sentences is easy to quote. A page that warms up for four paragraphs is not.
- Specificity over adjectives. "Payroll for Austrian construction firms with 5–50 staff" can be matched to a question. "Excellence in business services" cannot.
- Structure a parser can follow. Real headings, short paragraphs, lists where lists make sense, and schema markup where it applies. None of this is new advice; it simply matters more now.
- Corroboration. Your claim about yourself is one source. The same fact visible in a directory, a review profile and a trade listing is three.
- Freshness where freshness is relevant. Prices, hours and service lists that are visibly current get chosen over pages last touched in 2021.
What has not changed
The fundamentals still carry most of the weight. A fast, mobile-friendly site with clear pages per service, a claimed and active map listing, and reviews that people actually leave — these still decide most local outcomes. AI search does not replace that groundwork; it raises the cost of skipping it, because a thin footprint is now not just badly ranked but unquotable.
Ask an AI assistant the question a customer would actually ask — "who does payroll for small firms in Vienna" — and see who gets named. Then ask it directly about your company and see what it believes. It takes five minutes and it is the closest thing to a free audit of how you look to the systems doing the summarising.
The uncomfortable part is that you cannot buy your way into a citation the way you can buy an ad slot. What gets cited is a business that is easy to describe accurately. That is a content and operations problem more than a marketing one.
Worth a conversation?
If you want a view on how findable your business actually is — and what would move it — that is twenty minutes, not a project.
Book a call