← All media signals
Interface

Post-Search Discovery

The answer arrives whole and sourced, and the ten blue links — the open web we browsed — quietly fade to grey.

What is it?

Search used to hand back links to sift; now it hands back one assembled, cited answer, and browsing disappears. Discovery moves from search-and-click to ask-and-be-told, with a model standing between query and the web itself. Whether the model’s pick is fair, or just first, stays open.

Why now?

Answer engines removed the reason to browse: assembling a cited response from many pages became cheaper than ranking and serving them one at a time. Platform disintermediation means the layer that routed traffic now absorbs it and keeps the click. The web still gets crawled, just not visited.

Who’s building it?

Search platforms and model providers now build discovery into chat and answer surfaces, above the open web.

OpenAI · ChatGPTGoogle · AI ModePublishersGEO agenciesAd platforms
The full read
Why now, who is building it, where the market goes, and what stays unsettled.
Get weekly →
01 · The market
Who’s in it
Builders — answer engines and the model providers behind them.
Buyers — everyday searchers now typing questions straight into chat instead of search boxes.
Gatekeepers — platforms deciding which answer, and which source, gets shown.
What’s at stake
The features that made search useful are the ones answer engines absorb — ranking gives way to being chosen.
Publisher referral traffic keeps thinning as citation replaces the click.
Being the cited source now outweighs ranking on the page.
What’s unsettled
Whether citation ever pays publishers the way ranking once did.
Whether the open web stays worth crawling once nobody clicks through.
Which handful of models end up deciding what counts as the answer, and for whom.
02 · How the demand matureda progression, not a timeline
Trial
Users try the assembled answer instead of clicking through, out of curiosity.
Habit
The answer satisfies enough that clicking through starts to feel optional, then rare.
Dependence
Publishers adjust to being read by a model instead of a person, chasing citation over rank.
Category
Answer visibility becomes its own channel, budgeted and optimised for, separate from search rank.
03 · Where the market goesthe forks, not a prediction
→Consolidate or fragment — a handful of answer engines could settle into the new front door, or the model layer stays crowded and contested.
→Regulators could force answer engines to disclose or pay for the sources they quote.
→Publishers could route around the model with paid licensing deals, or lose the traffic outright.
→Discovery could stay hybrid, with links surviving for tasks answers can’t finish.
The open question is whether being cited replaces being visited as the real currency of discovery — or whether clicks still decide who gets paid.
04 · The Great Unbundling deskadjacent markets, moving now
Post-App Phones
Ask for outcomes, skip the app.
Apple’s 2026 Siri redesign lets users ask outcomes directly.
TechCrunch
The Model Mirror
Consumers now ask the model instead of searching.
50% of consumers now use AI search tools.
CMSWire (McKinsey research)
The Model Native Brand
Brands optimise to be quoted by models.
GEO can lift AI-answer visibility up to 40%.
arXiv (Princeton/Georgia Tech GEO study)
05 · The Wire
01
Google said AI Mode would reach all US users in 2025
TechCrunch ↗
02
Reuters Institute finds a third less Google traffic reaches publishers
Press Gazette ↗
03
Google says AI Mode now has over 1 billion monthly users
9to5Google ↗
The panel on this signal
They agreed on
–The three analysts agreed discovery is shifting from ranking pages to being the source a model quotes.
–Publishers now compete for citation, not page position.
–Whoever controls the model layer holds new leverage over who gets seen at all.
The platform analyst disagreed on
–Query complexity is the relevant variable: simple factual lookups may resolve fully within a generated answer, but higher-stakes or ambiguous searches still show users clicking through to source pages to verify before acting.
Know more about the method →
✦How this signal was built
01 · Panel
A Delphi panel of 4 AI analysts, each reasoning independently.
→
02 · Debate
They challenge and revise each other’s reads over two rounds.
→
03 · Evidence
Every claim is tied to a cited, linked source on The Wire.
→
04 · Synthesis
A moderator writes it into one read, disagreement and all.