The search box is a piece of furniture, and marketing strategy has been built around its shape for twenty years. Keyword research, page titles, meta descriptions and the whole apparatus of optimisation assume a person typing a short phrase into a rectangle. That assumption is now only partly true, and the part that has changed carries most of the strategic consequence.
Elliott King set out the shift at the BETT Higher Education Conference in two thousand and seventeen, closing a seminar delivered for the agency MintTwist with colleagues from MintTwist and City, University of London. The statistics in that room belonged to his colleague Colin Cheng. The framing at the end was his.
A spoken question is a different unit of demand
"They are probably connected to the internet in their sleep via their connected devices. The first thing most of them will do when they wake up is check their social media. We need to be AI ready. They are intelligent interfaces that allow us to speak to them and ask them questions in a semantic manner, and they will intelligently draw down answers to our questions. That is fundamentally different to typing a search into a search engine."
Watch this moment, 26:45
Two differences matter commercially. A spoken question is longer, and it carries its own qualifying detail: constraints, location and purpose arrive inside the query rather than being inferred from it. And a spoken question expects a single answer rather than a page of options, which removes the position-by-position competition that ranking strategy was designed around.
Search remains the most honest record of demand
Chapter 5, Search, treats the search engine as the clearest available statement of what a market actually wants, because a query is a declared intention rather than an expressed preference. That principle survives the change of interface intact. What changes is the shape of the record.
Short keyword phrases describe categories. Long spoken questions describe situations. A organisation that has built its content around the first will find it ranks for subjects while being absent from the moments where decisions are actually made, which is a more expensive absence than a lost position.
The top of the funnel has moved further from the brand
Chapter 2, Digital Marketing Strategy 101, describes how the digital age handed control of the upper funnel to the customer. A synthesised answer extends that transfer: the awareness phase can now complete without the customer visiting anything the organisation owns.
The practical consequence for the funnel is that awareness and engagement compress into a single exchange. A person asks, receives an answer assembled from published sources, and arrives at the owned estate already informed, already comparing, and frequently already decided. Content written to introduce a subject meets an audience that has been introduced to it by somebody else.
The response is not to abandon owned media. It is to recognise which of it is now doing awareness work at one remove: the material a machine can retrieve, weigh and summarise on the organisation's behalf, which is the same specific, attributable, well-structured writing that Chapter 7, Websites, asks for on other grounds.
Being retrievable is a content decision before it is a technical one
A page that cannot be fetched and parsed is invisible, and that part is a solved checklist. Beyond it the question becomes editorial. An answer engine draws on sources that state something specific, attribute it to somebody identifiable, and stand still long enough to be cited more than once.
Generic category content satisfies none of those conditions. It is the marketing equivalent of an unsigned document: technically present, practically unusable as evidence. Chapter 4, Digital Marketing Tools and Tactics, makes the case for episodic, expert-led content on engagement grounds. The retrieval argument reaches the same conclusion by a different route.
The measurement question changes with the interface
Position tracking answers a question the market is increasingly not asking. The useful measurement is whether the organisation appears in the answers its buyers actually receive, across the full spread of phrasings a spoken question can take, and whether what is said about it there is accurate.
That is a broader and less tidy instrument than a ranking report, and it is the one that corresponds to how the decision is now made. Chapter 9, Managing an Integrated Strategy, argues for measurement that a commercial team would recognise rather than a channel metric that flatters the discipline. The principle applies here without amendment.
Elliott King, co-author of this book and an AI visibility expert, has written about what this means for organisations in his own account of the same seminar. The book sets out the search and content frameworks, and his work at FINN Partners applies them.
Frequently asked questions
How is an AI answer different from a search result?
A search result offers a list of positions and leaves the comparison to the reader. An answer is synthesised from several sources and returned as prose, so the competition is for inclusion in the answer rather than for a position above a rival. Appearing at all becomes the binding question.
Does search engine optimisation still matter?
The technical foundation matters exactly as much, because a page that cannot be retrieved cannot be cited. What has diminished is position-by-position ranking strategy, and what has grown is whether the content states something specific enough and attributable enough to be drawn on as a source.
What should a marketing team measure instead of rankings?
Whether the organisation appears in the answers its buyers receive for the questions they actually ask, and whether the characterisation is accurate. That requires testing real question phrasings rather than keywords, and reading what comes back rather than counting a position.