Interactive model
Tap a stage to see how it feeds the next
Trust
01
Be Understood
Establish one unambiguous business entity — who you are, what you do, who you serve, where you operate and who is behind it.
Outcome: AI identifies you correctly and tells you apart from similarly named businesses. Feeds into Be the Answer.
At the centre: Machine Trust
Machine Trust is the degree to which AI systems can identify the business correctly, verify its claims, understand what it is known for, and confidently include it in an answer or recommendation.
Every stage either builds Machine Trust or converts it into a commercial outcome.
The evidence
Why the flywheel matters now: what the data says
Every figure below comes from the organisation that published it, linked so you can check it yourself. Where studies disagree, we say so.
Buyers moved to AI answers, fast
ChatGPT weekly active users, in millions, as announced by OpenAI. This is the audience Be Understood and Be the Answer are built for.
When AI answers, fewer people click
Share of Google visits that ended with a click on a regular result. Pew tracked the real browsing of 900 US adults in March 2025.
Ranking #1 is worth less than it was
Drop in click-through rate for the top-ranking page when an AI Overview appears. Ahrefs re-ran its study as AI Overviews expanded.
Being cited is the new ranking
Across 3,100+ queries (Jun 2024 to Sep 2025), Seer found AI Overviews cut organic CTR by 61%, yet sites cited inside them saw higher CTR. Seer notes this is correlation, not proven cause. Its 2026 update found click-through rates are levelling off, not recovering.
+35%
higher organic click-through rate when cited in the AI Overview
+91%
higher paid click-through rate when cited in the AI Overview
What this means: the click is shrinking, the citation is not. The flywheel is built to win the citation, then the recommendation.
A note on honesty: Gartner's 25% figure is a forecast, and some analysts, including Search Engine Journal, dispute it. We show it as a prediction, not a fact.
The six stages
Stage 1
Be Understood
Establish one unambiguous business entity
Before AI can recommend a business, it must understand who the business is, what it does, who it serves, where it operates, who is behind it, and which profiles and mentions belong to the same organisation.
Core work
- •Canonical business identity
- •Entity home
- •Clear About page
- •Organization and Person schema
- •sameAs connections
- •NAP consistency
- •Service definitions
- •Founder and expert profiles
- •Knowledge-graph corroboration
Desired outcome: AI can identify the business correctly and distinguish it from similarly named organisations.
Feeds into Be the Answer
Stage 2
Be the Answer
Publish information AI can lift and reuse
Once AI understands the entity, the business must publish useful answers to the questions customers actually ask.
Core work
- •Question-shaped headings
- •Answer-first paragraphs
- •Extractable claims
- •Self-contained content sections
- •Buying guides
- •Service explanations
- •Comparison pages
- •FAQs
- •Tables
- •Pros and cons
- •Original insights
- •Clear evidence beneath every claim
Desired outcome: When AI searches for an answer, the business has a clear, complete and relevant passage available.
Feeds into Be Cited
Stage 3
Be Cited
Build the architecture that earns attribution
Being the answer is not enough. The information must appear in sources and formats AI systems trust enough to cite.
Core work
- •Citation seeding
- •Original research and data
- •Digital PR
- •Relevant directories
- •Industry publications
- •Expert commentary
- •Podcasts
- •Professional associations
- •Review platforms
- •Community participation
- •Source-authority mapping
- •Consistent third-party mentions
Desired outcome: AI systems use the business or its content as supporting evidence inside generated answers.
Feeds into Be Recommended
Stage 4
Be Recommended
Move from being a source to being a named choice
A business may be cited for information without being recommended as the provider. This stage focuses on winning specific commercial comparison contexts.
Core work
- •Identify valuable “best for” territories
- •Map recommendation prompts
- •Map comparison prompts
- •Build co-occurrence between the brand and its specialty
- •Strengthen review depth and specificity
- •Build consensus across independent sources
- •Publish “why choose,” comparison and alternatives content
- •Earn inclusion in relevant “best of” resources
Desired outcome: When a potential customer asks who they should choose, the business is named as a relevant option.
Feeds into Be Chosen
Stage 5
Be Chosen
Convert AI-referred buyers into customers
Being recommended does not automatically mean being selected. The website and sales journey must confirm the recommendation rather than forcing the prospect to restart their research.
Core work
- •Recommendation-matched landing pages
- •Clear proof and positioning
- •Shorter buying journeys
- •Trust-assumptive sales experiences
- •Relevant case studies
- •Clear comparisons
- •Low-friction booking or purchasing
- •Attribution capture
- •“How did you hear about us?” with AI as an explicit option
Desired outcome: A prospect arrives already informed and pre-sold, and the business makes it easy for them to act.
Feeds into Be Proven
Stage 6
Be Proven
Turn every customer result into fresh evidence
Every new customer can create stronger proof that feeds the next turn of the flywheel.
Core work
- •Specific customer reviews
- •Detailed case studies
- •Measurable outcomes
- •Customer stories
- •Third-party mentions
- •Before-and-after evidence
- •Updated statistics
- •Freshness signals
- •New expert insights
- •Recommendation-win screenshots
- •Ongoing visibility measurement
Desired outcome: Real customer outcomes become machine-visible evidence that strengthens the business’s identity, credibility and future recommendations.
Feeds into Be Understood (the next turn)
Why it is a flywheel, not a funnel
A funnel ends when the customer buys. The AI Search Flywheel continues: the output of the final stage becomes an input to the first.
Clearer identity creates better answers. Better answers earn citations. Citations build recommendations. Recommendations create customers. Customers create proof. Proof strengthens identity.
Quick reference
- Be Understood
- Establish an unambiguous entity so machines can resolve who you are.
- Be the Answer
- Publish answer-ready content machines can lift whole.
- Be Cited
- Build the citation architecture that gets your claims credited inside AI answers.
- Be Recommended
- Graduate from cited source to named suggestion inside chosen comparison contexts.
- Be Chosen
- Convert AI-referred, pre-sold buyers without re-selling them from zero.
- Be Proven
- Turn every customer into fresh machine-visible evidence that feeds the flywheel again.