For twenty years, search engine optimisation meant one thing: get your page into the top handful of blue links. That objective quietly stopped being sufficient somewhere around 2025, and in 2026 it is actively misleading. The question is no longer whether you rank. It is whether you get cited.
The zero-click problem is not a blip
Roughly six in ten searches now conclude without anyone clicking through to a website. The answer gets assembled on the results page itself, drawn from sources the user never visits. For informational queries — the “how does X work”, “what is Y” content most businesses publish — the AI summary now resolves the question before organic results get a look in.
It is worth being precise about what this does and doesn't change. Commercial and transactional searches still behave much as they always have. When somebody is comparing providers, requesting a quote or about to spend money, the top organic position still earns the click, because a synthesised paragraph cannot complete the transaction. The damage is concentrated in top-of-funnel informational content.
The practical read
If your traffic model depends on informational blog content pulling visitors who later convert, that funnel is leaking at the top and no amount of traditional keyword work will re-seal it. If your traffic is mostly commercial-intent, you have more time than the headlines suggest — but not indefinitely.
GEO, AEO, and the acronym problem
The industry has produced an unhelpful cloud of terminology — Generative Engine Optimisation, Answer Engine Optimisation, LLM Optimisation, AI Optimisation. Strip the labels away and they describe the same discipline: structuring content so that a machine can extract it, verify it, and reproduce it accurately with attribution.
That is a meaningfully different craft from keyword targeting. A language model assembling an answer is not scanning for exact-match phrases. It is looking for claims it can lift with confidence. Content that buries its answer in the eighth paragraph of a meandering introduction is content that will not be quoted, regardless of how well it is keyword-optimised.
What actually moves the needle
1. Answer the question in the first paragraph
Lead with the conclusion, then justify it. This is the inverse of the SEO-era house style, where writers padded the opening to delay the answer and inflate time-on-page. That tactic now costs you citations.
2. Structured data has stopped being optional
Schema markup — Article, FAQPage, HowTo, Organization — is the most direct way to tell a machine what a page is, who wrote it, and what it covers. It remains one of the most consistently underused tools available, largely because it delivers nothing visible on the page and therefore never feels urgent.
3. E-E-A-T is now a filter, not a ranking factor
Experience, Expertise, Authoritativeness and Trustworthiness used to nudge you up or down a results page. In an AI-mediated search environment the function has changed: if a system does not trust your brand, you are not ranked lower, you are simply omitted from the answer. There is no page two to fall back to.
4. Search is no longer only Google
It bears repeating that the scale gap remains enormous — Google still processes billions of queries a day against a few tens of millions for the largest AI assistants. Anyone claiming Google is finished is selling something. But referral traffic from generative AI has grown by an order of magnitude in a short window, and buyers increasingly consult an assistant to shortlist vendors before they ever open a search engine. Being absent from that shortlist is expensive in a way that is very hard to measure.
The uncomfortable part
Most of this work is unglamorous. Clean HTML structure. Accurate, specific claims. Author attribution that means something. Schema that reflects reality. Content with a genuine point of view that a summary cannot flatten. It looks a great deal like doing the job properly, which is precisely why so many sites will not do it.
Key takeaways
- Zero-click is structural, not a phase — plan traffic assumptions around it
- Commercial-intent pages still earn clicks; informational pages are where the loss is
- Lead with your answer; AI systems cite extractable claims, not slow build-ups
- Schema markup is the cheapest credibility signal you are probably ignoring
- Trust now decides inclusion, not position — there is no second page in an AI answer
Want this handled properly?
Tell us what you're working with and we'll scope it with you — no obligation.