Search Foundations in the AI/Zero-Click Era

Overview: the shift this course encodes

v1 of this course taught ranking. v2 teaches visibility and return, of which ranking is one component. The practical consequence is that you will evaluate every content idea twice: once for whether it can rank, and once for whether it can survive being answered by a machine.

1.1 What is modern search?

What is it?

Modern search is a portfolio of discovery surfaces. A single commercial question may be answered by an AI summary at the top of a results page, a classic organic result, a video, a forum thread, a map pack, a shopping module, a vertical marketplace, or a conversation inside a standalone AI assistant that never shows a results page at all.

Why it matters

If you optimise for one surface you inherit its volatility. Publishers who treat "position 3 on desktop Google" as the product are exposed to a layout change they cannot influence. Publishers who know which surfaces their buyer actually uses can spend where clicks still exist.

How to do it

1. Choose ten real queries your buyer would type or say, spread across informational, commercial and transactional intent.
2. For each, record what appears: AI summary, ads, classic results, video, forums, maps, shopping, "people also ask", vertical results.
3. Note where the first classic organic result actually sits on screen — count scroll depth, not rank number.
4. Repeat on mobile. The surface mix differs enough to change strategy.
5. Ask the same questions of two AI assistants and record whether any source is cited, and which.
6. Mark each query: click-rich, click-poor or click-dead.

Tools needed

A browser in a clean profile, one mobile device, two AI assistants, and a spreadsheet. No paid tools required for this lesson.

Lab 1.1

Build the surface inventory for ten queries in SMB Finance and Banking. Include at least two queries you personally would want to rank for. Conclusion required: which three queries are still worth a page, and why.

Artifact

Surface inventory
Query | Intent | Surfaces present | Scroll depth to first organic |
Cited sources in AI answer | Verdict (click-rich / click-poor / click-dead)

Common mistakes

• Auditing from a personalised, logged-in browser and mistaking your own history for the market.
• Ignoring forums and video, which frequently outrank publisher content for experience-led queries.
• Recording rank position without recording what is above it.

Pro astuces

• Re-run the inventory quarterly and diff it. Surface change is your leading indicator of traffic change.
• Photograph the whole viewport, not the result list — you are documenting attention, not rank.

1.2 What is zero-click search?

What is it?

A zero-click search is one where the user's need is satisfied on the results surface — by a summary, a definition, a snippet, a calculator widget, a map or an assistant's answer — so no publisher receives a visit.

Why it matters

Zero-click behaviour does not hit all content equally. It removes demand for content that can be finished in a paragraph, and leaves demand for content that requires personalisation, current data, interaction, judgement, or a relationship. That asymmetry is the entire strategy of this academy.

How to do it

1. Take every planned content idea and run it through the seven survival questions below.
2. Score each question 0 or 1; total of 0–2 means the idea is displacement bait.
3. For low scorers, ask: can this become a calculator, template, tracker or comparison instead?
4. Kill anything that cannot be upgraded and does not support a tool, money page, data asset or the newsletter.
5. Record the decision so you do not re-litigate it in three months.

Zero-click survival scorecard
Can AI fully answer this?                   (invert: 1 if no)
Does the page contain original data?
Does the page contain a tool/template?
Does the user need personalization?
Does the user need updated information?
Does the page create an email/signup reason?
Would someone bookmark this?
-------------------------------------------------
Score /7      Verdict: build / upgrade / drop

Tools needed

The zero-click survival scorecard tool, your surface inventory from 1.1, and two AI assistants for the "can it be answered" test.

Lab 1.2

Score twelve candidate assets for one Phase 1 site — deliberately include four generic explainers. Upgrade or drop every idea scoring 2 or below and show the upgraded version. You should end with a list that resembles the 12-asset MVP: 2 pillars, 3 money, 2 tools, 2 data, 3 support.

Artifact

Zero-click survival scorecard for 12 assets, with upgrade decisions

Common mistakes

• Assuming volume equals opportunity. High-volume definitional queries are the most displaced.
• "Upgrading" a thin explainer by adding words rather than adding a tool, data or judgement.
• Writing for the query instead of the job the query is a symptom of.

Pro astuces

• Ask an assistant your target question and read its answer as your competitor's brief. Publish what it cannot do.
• Any asset scoring 6–7 deserves a bigger budget than three articles combined.
• Dated assets ("2026 fee tracker") buy you recurring demand — but budget the refresh forever.

1.3 What are crawling, indexing, ranking, retrieval, and citation?

What is it?

Five distinct stages, often collapsed into the word "SEO":

Why it matters

Ranking is no longer the only visibility mechanism, and it is possible to be retrieved without being cited, or cited without being clicked. Diagnosis fails when you treat one funnel as five. A page that is not indexed has a technical problem; a page indexed but never retrieved has a content-structure problem; a page retrieved but never cited usually has an originality problem.

How to do it

1. Confirm fetchability: request the URL as a bot would, check status code and rendered HTML.
2. Confirm indexing in Search Console URL inspection, not by a site: search alone.
3. Check ranking for its target query set and record the surfaces present.
4. Test retrieval: ask assistants the question the page answers, and see whether anything resembling your content appears.
5. Test citation: check whether you or a competitor is named, and what makes theirs citable.
6. Log one fix per failing stage — never mix stages in a single ticket.

Tools needed

Google Search Console, a crawler, browser dev tools, and two AI assistants for the retrieval and citation checks.

Lab 1.3

Take one existing page (yours or a competitor's) and complete a five-stage audit with evidence at each stage. Then write the single highest-leverage fix.

Artifact

Retrieval readiness audit
URL | Crawlable? | Indexed? | Ranking? | Retrieved? | Cited? |
Failing stage | Root cause | One fix | Owner | Recheck date

Common mistakes

• Blaming the algorithm when the page is blocked, canonicalised away or duplicated.
• Rewriting content to fix an indexing problem, or fixing tags to fix an originality problem.
• Treating one assistant's answer on one day as a measurement. Sample repeatedly and log dates.

Pro astuces

• Publish a methodology section on data assets. It gives a generated answer something concrete to attribute.
• Keep URLs stable. Citation accrues to addresses, and redirect churn resets it.

1.4 What is search intent vs task intent vs return intent?

What is it?

• Search intent — what the person typed and the immediate answer type they expect.
• Task intent — the job they are trying to complete this week (open an account, choose a vendor, pass an audit).
• Return intent — the reason they would come back next month without searching.

Why it matters

Search intent alone produces pages that answer and end. Task intent produces tools and templates. Return intent produces newsletters, trackers and communities — which is what makes traffic survivable. Every asset in the 12-asset MVP should serve at least two of the three.

How to do it

1. Write the query. Then write the sentence the user would say to a colleague about what they are actually doing.
2. Identify the deliverable that finishes the task (a number, a shortlist, a filled template, a decision).
3. Design the on-page element that produces that deliverable.
4. Identify the recurring trigger — a rate change, a renewal, a deadline, a quarterly report.
5. Attach the return mechanism: newsletter segment, tracker subscription, saved calculation.

Three-intent map
Query | Search intent | Task intent | Deliverable |
Return trigger | Return mechanism | Asset type | Priority

Tools needed

Your surface inventory, three real conversations with people in the audience, and the content brief builder.

Lab 1.4

Map all three intents for the twelve MVP assets of one Phase 1 site. Any asset with no return trigger must either gain one or be justified as pure acquisition.

Artifact

Three-intent map for 12 assets

Common mistakes

• Inventing task intent from imagination instead of asking three real people.
• Bolting a newsletter box onto a page with no recurring reason to subscribe.
• Serving return intent only through email while the site itself never changes.

Pro astuces

• The best lead magnet is usually the deliverable of the task, saved and emailed.
• Deadline-driven niches (tax, renewals, enrolment) have return intent built in — exploit the calendar.

1.5 What is topical authority plus brand demand?

What is it?

Topical authority is depth and coherence across a subject. Brand demand is people asking for you — branded searches, direct visits, newsletter opens, repeat sessions, mentions. v2 treats brand demand as the headline KPI and topical authority as the means.

Why it matters

Topical authority is judged by systems you do not control. Brand demand is an asset you own; it converts better, it is not reallocated by a layout change, and it is what an acquirer pays for.

How to do it

1. Define the subject boundary in one sentence — what you will and will not cover.
2. Build coherent clusters around the 12 MVP assets, not scattered keyword pages.
3. Baseline brand metrics now: branded impressions, direct sessions, returning users, subscribers.
4. Give people a name to remember: the tracker, the calculator, the weekly brief.
5. Report brand demand monthly next to organic clicks, in the same table.

Brand demand baseline
Branded impressions/clicks (GSC)   Newsletter subscribers
Direct sessions                    Open + click rate
Returning users %                  Tool completions
Email-sourced sessions             Unlinked brand mentions

Tools needed

Search Console, GA4 (or Cloudflare Web Analytics), your newsletter platform, and a monthly reporting sheet.

Lab 1.5

Record today's baseline for all eight metrics, even if seven of them are zero. Set the month 6 target for each and note which single asset you expect to move them.

Artifact

Brand demand baseline with month 6 targets

Common mistakes

• Publishing broadly to "cover the topic" and diluting the boundary.
• Measuring brand demand only after growth stalls, so you have no baseline to compare against.
• Naming nothing. Unnamed assets do not generate branded searches.

Pro astuces

• Name your recurring assets like products: "the Business Banking Fee Tracker", not "our fees page".
• Watch the ratio of branded to non-branded impressions over time; rising ratio means durability.

1.6 What is trust in YMYL and AI-mediated discovery?

What is it?

Trust is the evidence a reader — or a system summarising for a reader — can use to decide whether to rely on you. In practice: E-E-A-T (experience, expertise, authoritativeness, trustworthiness) made visible through named authors, verified reviewers, sources, methodology, ownership disclosure and update history.

Why it matters

Money-or-life topics — finance, insurance, health, legal, safety — carry a higher evidence bar because bad information causes real harm. In AI-mediated discovery, the same visible signals help a system decide whether your page is a safe thing to summarise and attribute.

How to do it

1. Put a real named author with a real bio and a real credential on every YMYL page.
2. Add a named reviewer where the topic requires professional judgement, with the review date.
3. Cite primary sources — regulators, official documentation, standards bodies — not aggregator blogs.
4. Publish a methodology page explaining how you test, score and compare.
5. Disclose ownership, funding, affiliate relationships and any common ownership across your sites.
6. Show verified and updated dates, and keep a public corrections log.

Trust surface audit
Named author with credential?      Corrections policy?
Named reviewer + date?             Ownership disclosure?
Primary sources cited?             Advertiser disclosure?
Methodology page?                  Contact route to a human?
Verified/updated dates?            Common-ownership statement?

Tools needed

Your reviewer register, an author bio template, and the review methodology page pattern.

Lab 1.6

Audit the ten trust surfaces on two competitors in a Phase 1 niche, then on your own plan. List the three you can implement before publishing your first asset.

Artifact

Trust surface audit + your three pre-launch trust commitments

Common mistakes

• An "Editorial team" byline with no human behind it.
• A reviewer name with no scope, no date and no contract.
• Disclosures buried in the footer while the affiliate links sit in the first paragraph.

Pro astuces

• Publish the reviewer's scope ("reviewed the tax treatment section") rather than implying they wrote it.
• A visible corrections log is cheap and disproportionately increases perceived trustworthiness.

Scenario assessment: the HSA vs FSA collapse

A page ranking in the top three for "HSA vs FSA" loses 55% of its clicks over four months after AI summaries begin appearing for the query. Impressions are flat. Average position is unchanged. The page earns 4% of site revenue, has one affiliate link, and took an expert reviewer four hours.

Final project: the zero-click survival pack

Deliverables
1. Surface inventory for 10 queries (desktop + mobile + 2 assistants)
2. Zero-click survival scorecard for 12 candidate assets
3. Retrieval readiness audit for 1 page across all 5 stages
4. Three-intent map for the 12 MVP assets
5. Brand demand baseline with month 6 targets
6. Trust surface audit with 3 pre-launch commitments

Pass standard: at least eight of your twelve assets score 4+ on the survival scorecard, every asset has a named return trigger, and no asset relies on a definition as its main value.

Course 1 checklist

• Surface inventory built from a clean, signed-out browser on desktop and mobile
• Every query classified click-rich, click-poor or click-dead
• 12 assets scored on the zero-click survival scorecard
• Every low scorer upgraded to a tool, dataset or comparison — or dropped
• Five-stage retrieval audit completed with one fix per failing stage
• Task intent validated with three real conversations, not assumptions
• Return trigger and mechanism defined for every asset
• Brand demand baseline recorded with month 6 targets
• Trust surfaces audited; author, reviewer and disclosure plan agreed
• No fabricated authors, reviews, screenshots or credentials anywhere in the plan

AI-agent prompt for Course 1

Act as a search strategist working in an AI-mediated, zero-click environment.

I will paste: my niche, my buyer, and a list of candidate content ideas.

Do this:
1. For each idea, score the seven zero-click survival questions (can AI fully answer it; original data; tool/template; personalization needed; freshness needed; email/signup reason; bookmark-worthy). Output a table with a score out of 7 and a verdict of build, upgrade or drop.
2. For every "upgrade" verdict, propose the specific tool, dataset, tracker or comparison that replaces the explainer, and name the deliverable the user walks away with.
3. Map search intent, task intent and return intent for the surviving ideas, including the recurring trigger that brings the user back.
4. Reduce the list to a 12-asset MVP: 2 pillar guides, 3 money pages, 2 tools/templates, 2 original data or tracker assets, 3 support articles.
5. For each asset, state which of crawling, indexing, ranking, retrieval or citation is the likely bottleneck and the first fix.
6. List the trust surfaces required before publishing in this niche (author, reviewer, sources, methodology, disclosures, dates).

Rules:
- Do not invent statistics, sources, reviewers or platform behaviour. Mark unknowns as "verify against primary source" with the source type.
- Never propose scaled templated pages, doorway patterns or fabricated testing.
- Prefer assets a reader would bookmark or subscribe to over assets that merely answer.
Output as tables.

Primary sources for this course

• Google Search Essentials
• Creating helpful, reliable, people-first content
• Google spam policies
• Google Search Console