Guides
AI is rewriting search, and what it means to be findable
Search is becoming an answer rather than a list of links. That changes what it means to be discoverable, and it puts a premium on having something that can be asked directly.
For twenty-five years, being findable meant ranking. Someone typed a query, got ten links, and clicked one. The entire discipline of search optimisation existed to influence which links appeared and in what order.
That model is being replaced by one where the search engine answers the question itself, citing sources, and the click is optional. The consequence for anyone whose work depends on being found is not subtle.
What changes for the person being searched for
Under the old model, your material competed for a position in a list. Under the new one, it competes to be the source an answer is synthesised from.
Three practical differences:
- Fewer visits, more mentions. Your material can shape an answer that never sends anyone to your page.
- Clarity beats volume. Systems synthesising answers reward material that states things plainly. Padding, which used to help ranking, now actively hurts extraction.
- Structure matters more. Clear headings, direct answers to specific questions and unambiguous claims are what gets pulled into an answer.
What does not change
The underlying need. Someone still has a question, and they still want a trustworthy answer from someone who knows.
What is changing is only the mechanism by which they get to it, and mechanisms have changed before. The people who did badly in previous transitions were the ones optimising for the mechanism rather than for the question.
The thing that survives every version of this
An interface that answers questions directly.
If a prospective client can ask something and get a specific, grounded answer from your material, that works whether they arrived from a search result, an AI-generated summary, a link a colleague sent, or a QR code on a conference badge. It does not depend on which retrieval system is currently dominant.
This is why an AI Twin is a reasonable hedge as much as a product. Whatever front end people use to look for expertise, the thing they eventually want is to ask a question and get your answer. Having that available as a durable, addressable thing is independent of the channel.
The practical implication
Two things worth doing regardless of what search becomes:
- Make your material extractable. Clear structure, direct claims, plain answers to the questions you actually get asked.
- Make it askable. Give people somewhere to put the follow-up question, and make sure the answer comes from your material rather than a general model guessing.
The first is table stakes and everyone is doing it. The second is where the difference currently is.
This piece revisits earlier essays from our Medium archive: Game of search and GPT tech will alter the user experience forever.
See what an AI Twin is.
AI Twin produces answers generated by AI. Important decisions should still be verified against the cited sources or with the expert directly.
