FM

Find A Masters AI Search Visualiser

Agentic RAG, pairwise ranking, and citation survival for postgraduate search.
The one-minute version

Old search picked a winner. AI search builds a case.

Take a question people actually argue about — watch how the two approaches answer it, then what that means for your content.

Plays the search out step-by-step
"Who's the best goalkeeper in the Premier League?"
The old way

Follow one signal and stop. "Arsenal won the league — so their keeper must be the best."

≈ one search, one hop — authority by association
What AI search does

Breaks the question into smaller searches — each one a real query it runs, routed to wherever that evidence lives. Your page only enters the answer if it's what one of those searches pulls in:

Save % — how many shots does he stop?
searches “best PL goalkeeper save % 2025/26”pulls in stats sites
Penalty saves, clean sheets, errors
searches “PL goalkeeper clean sheets & errors 2025/26”pulls in match reports
Does he pass the eye test?
searches “best goalkeeper saves this season”pulls in YouTube highlights
What do fans actually think?
searches “best Premier League keeper right now reddit”pulls in Reddit & X threads
Is he his country's No. 1?
searches “is he England's first-choice goalkeeper”pulls in squad news — independent authority
≈ query fan-out into real sub-queries → each routed to a retrieval tool → passages (maybe yours) pulled in

Collate & compare. Pool every source, then weigh the leading keepers head-to-head. Strong across many measures beats a single freak stat.

≈ reciprocal rank fusion → pairwise ranking

The verdict — quoting its sources. "He leads on goals prevented in the data, the highlights back it up, and the fans agree" — one answer stitched from several sources, each named.

≈ synthesis with citations
Now flip it — that's you. Each of those searches is a door your content can walk through: be the stats site and you're pulled in on the save-% search; be the explainer and you're pulled in on a "how does it work" search. You don't earn a citation by ranking #1 for the big question — you earn it by being what one of the small searches turns up. And because one answer runs many searches, a single citation hides most of the story: classic rank tracking misses around 75% of your real footprint.

None of this is hand-waving: under the hood it's a documented pipeline — fan-out, routing, rank fusion, pairwise ranking, synthesis — each backed by a Google patent. The "Explore the mechanics ▸" tabs show every step.

The same shift, in plain terms

Search stopped being a ranking. It became a research assistant.

The abstract version of the story above — kept here while we decide which framing lands better.

The old way

You type a search. Google shows ten blue links.

Being #1 won the click.

What happens now

AI splits your question into smaller ones, searches a different place for each, compares what it finds, and writes a single answer — quoting a handful of sources.

Getting quoted in the answer is the new win.

So your content no longer just needs to rank — it needs to survive long enough to get quoted. And because one answer pulls from many searches, a single citation hides most of the story: appear in 4 of 12 sub-searches but get cited once, and classic rank tracking misses around 75% of your real footprint.

Why content gets left out

To be quoted, a page clears five gates in a row

Miss any one and you're not in the answer. This is a map of where content can fall out — not a literal look inside Google.

1The splitDoes your topic appear in any of the smaller questions Google generates?
2The surfaceBlog, video, shop, forum — does Google look somewhere you actually exist?
3The shortlistDoes your page get pulled into the candidates for that search? (Video and forum posts can get here without ranking #1.)
4The face-offRead next to a rival page, does yours answer the question better?
5The checkFresh, consistent, and specific enough to be worth quoting?
One real example

Walk one question through the five gates

A searcher asks "—". Here's the path to the answer — frozen, nothing to click.

So what do we actually do

Five priorities that decide whether you get cited

No guarantee — but content that scores well across all five tends to get quoted. This is the working checklist the rest of the report turns into real actions.

1

Coverage

Be present across the sub-questions, not just the headline term.

2

Format

Cover the right surface for each intent — guide, video, forum, product.

3

Evidence

Make passages specific enough to quote — names, numbers, real steps.

4

Comparison

Beat the competing page when an AI reads both side by side.

5

Action

Decide what to create or fix — and give it an owner.

Go deeper

See the machinery — or see where you stand

The tabs below are optional depth: each steps through one stage at your own pace. When you're ready for your real data, jump to the report.

AI search works more like a research assistant than a ranking engine.

It breaks a question into smaller jobs, searches different places for each job, compares the evidence it finds, and builds an answer from the strongest passages. Your content doesn't just need to rank. It needs to be discoverable across the right subtopics, available in the right formats, specific enough to beat competing passages, and useful enough to be cited in the final answer.

We can't see inside Google. This is the best working model based on Google's patents and observed behaviour — useful as a "where can content fail to be cited" map, not a literal architecture.

In practice, your page has to clear five gates in a row to be cited:

1The splitDoes your topic even show up in any of the smaller questions Google generates?
2The routeDoes Google pick a place you exist (blog, video, shop, Reddit, API)?
3The shortlistDoes your page get pulled into the candidates for that search? (Video and forum posts can get here without ranking #1.)
4The face-offWhen the AI reads your page next to a rival's and compares them, does yours win?
5The checkIs your page fresh, non-contradictory, and the kind of source the AI wants to quote?
A single citation under-reports your real footprint by 3–10×. Show up in 4 of 12 sub-retrievals, get cited once, and classic tracking misses 75% of your impact. That is why this report measures sub-query coverage, survival rate, and citation share — not rank.
Scenario Control

Choose a scenario

Pick a query and step through the agentic-search loop one stage per click. The scenarios are illustrative — synthetic examples chosen to show what a planner, router, retriever, critic and pairwise referee would each do. For your real citation data on these topics, see the View 2 Audit.

What you'll see · 1

The sub-queries

  • Google's smaller questions will list here.
What you'll see · 2

What each search returned

  • Top-10 results from each Google search will list here.
What you'll see · 3

The AI's verdict

After the AI reads the candidates head-to-head, its judgment lands here.

Live Execution Pipeline

The workflow: plan, search, judge. These are stages of one model — not separate agents.
Idle
Step 1 · User asks ↓ —
Intent reading · what the user really means
Step 2 · Agentic orchestration
Plan Decompose Split the question into focused sub-queries
Judge Critique & rank Catch gaps; pairwise head-to-head per role
Search Route & retrieve Pick the right surface, fetch candidate passages
Tools the agent reaches
Blog
Shop / product
Forum
YouTube / short
Vector index
Live web fetch
MCP / product API
Quality check · freshness, contradictions, source diversity
Step 3 · Returns to the searcher ↑
Run the simulation to see how the agent stitches a composite answer from multiple surviving passages.
What this view shows

The AI picks 3–5 pages by comparing them, not by ranking

After Google ranks the top 10 for a search, the AI doesn't just pick #1. It reads two pages at a time and asks itself: "Which one better answers this question?" Across many pairs, it adds up the wins. The pages with the most wins get cited in the AI Overview.

Below, we walk through this for one smaller question: "—" from the scenario "—". Four pages came back from Google — watch them fight it out.

View mode
How candidates get to the AI

Google ranks the top 10. Then the AI reads them.

Google's normal ranking picks the top 10 for the sub-query. The AI then reads the actual page text of those 10 and compares them two at a time. If your page doesn't make Google's top 10, the AI never reads it.

1 · The sub-query
"—"
A literal Google search runs for this text.
Google
ranks it
2 · Top 10
10 results
The SERP. Same retrieval classic SEO has always targeted.
AI
compares
3 · Pairwise gate
4 candidates
The AI compares these head-to-head. Aggregate W-L decides what gets cited.

The 4 candidates · with live W-L tally

Ready to judge

Round 0 of 4
Page A

Press play below — one round per click.

Page B

The AI reads both, then picks.

AI's verdict
After each round, the verdict + the deciding phrase land here.
What this view shows

One question becomes many — then gets stitched back into one answer

A single user question gets split into 5 smaller questions. Each one runs as its own Google search — some on the regular web, some on YouTube, some on Reddit. The results get combined into one list, the AI compares the top pages, and picks 3–5 to cite in the final answer.

Google doesn't publish exactly how it does this — what's below is our best read, grounded in the patents Google has filed. See docs/research/agentic-rag.md for what's confirmed vs inferred.

Question we're simulating: "—"

Three generations of how AI search works. The framework — sub-query coverage, survival rate, citation share — holds across all three. What changes is whether the stages are visible as separate steps or fused into one model.

2022 · old way

One search, one answer

The model runs one search, grabs a handful of nearby pages, and writes from those. No splitting, no comparing, no second pass.

  • Weak on questions that have several parts.
  • Doesn't route to forums, video, or specialised data.
  • No checking that the evidence is good enough.
  • Generic copy slips through because pages aren't compared.
2024 · today

Many searches, then judged

The system splits the question, runs many searches, compares the top pages two at a time, then writes the answer. This is what the rest of this report visualises.

  • Splits the question into 5–20 smaller related questions.
  • Routes each to its best source — blog, forum, video, shop.
  • Goes back for more if the evidence is thin.
  • Picks pages by comparing them, not by ranking signals alone.
2026+ · what's next

One model does it all

A single model trained specifically for search — splitting, retrieving, comparing, and writing all happen inside it. Domain-tuned versions are emerging (Charcoal, SID-1, Waldo).

  • No separate splitter / router / critic — the model owns the whole loop.
  • Tunable to a specific domain or vocabulary.
  • Better at the "last 20%" — your jargon, your aliases.
  • Risk: citations may fragment across many search systems.
Where to focus content effort

Five priorities that decide whether your content gets cited

AI search doesn't reward one #1 page anymore — it rewards content that meets a higher bar across these five dimensions. Treat this as a working checklist that emerges from the model we just walked through.

No guarantee of citation — but pages that score well across all five tend to. Use it to prioritise where to invest content effort.

1

Coverage

Be present across the sub-questions Google is generating, not just the headline query. Map your topic out to the 5–10 related angles a user would ask next.

2

Format

Cover the right surface for each intent — guide, video, forum, product page, FAQ. Google routes different sub-queries to different surfaces; if you only exist in one, you only reach one.

3

Evidence

Make passages specific enough to quote — named entities, concrete steps, real numbers, real comparisons. Generic framing loses the LLM-as-judge comparison every time.

4

Comparison

Beat the competing passage when an LLM reads both side-by-side. Specificity and direct-answer phrasing tend to win. Audit your top-cited competitors — what would beat them?

5

Action

Decide what to create or improve to lift the previous four — and assign an owner. The framework only works if it becomes a roadmap, not a poster.