Skip to main content Scroll Top

How AI Search Works: The Eight Steps From Question to Citation

Updated September 2026 · Written and maintained by the Progression Agency strategy team

The pipeline behind a cited answer — which steps you can influence, which are closed to everyone, and why a thinner competitor gets quoted instead of your better page.

On this page · 14 sections
  1. How AI search actually works
  2. Retrieval-augmented generation, without the jargon
  3. Why retrieval selects passages, not pages
  4. How the major surfaces differ mechanically
  5. Why your new page has not appeared yet
  6. What is documented, what is observed, and what is inference
  7. Where this leaves the practical work
  8. A note on which query types this mechanism favours
  9. Everything else we have written on search, AI and getting found
  10. The eight steps, and who controls each one
  11. The same question, four different machines
  12. Failure modes, mapped to the step they break
  13. How we know what we know
  14. Video: AI search, retrieval and how answers get built

The short answerEight steps run between a typed question and a named source: the question arrives, the system decides whether to search at all, it rewrites your question into queries you never see, retrieves candidates, extracts passages rather than pages, ranks them on relevance and apparent reliability, composes an answer, and names three or four sources. You can influence retrieval, extraction and selection. You cannot influence whether it searches, what it searches for, or which sources it finally names — which is precisely what guaranteed-citation offers claim to control.

Claims on this page are labelled by evidence class: documented by providers, observable by testing, or inference. Ranking weights and selection reasoning are published by nobody.

The pipeline, end to end
AEO can influence steps four, five and six. Step two decides whether you get a chance at all, and step eight is closed to everybody.

How AI search actually works

Eight steps between a typed question and a named source. Most writing on this subject skips straight to advice; the advice only makes sense once you can see which steps are addressable and which are closed.

The short version: the system decides whether to search at all, rewrites your question into queries you never see, retrieves candidate documents, extracts passages rather than pages, ranks those passages on relevance and apparent reliability, writes an answer from whichever survive, and names three or four sources.

You can influence retrieval, extraction and selection. You cannot influence whether it searches, what queries it generates, or which sources it finally names. That last one is worth sitting with, because it is exactly what guaranteed-citation offers claim to control.

Step 4 — Retrieval. Can your page be found as a candidate at all?.
Step 5 — Extraction. Is there a clean passage to lift?.
Step 6 — Selection. Does it look relevant and reliable?.
Step 2 — Search decision. Not yours. The system decides if it searches.
Step 3 — Query rewriting. Not yours. You never see the generated queries.
Step 8 — Attribution. Not yours, not anyone's, not for sale.
Which steps you can actually influence
Three reds. The most commonly sold promise in this field — guaranteed citations — sits squarely on the last one.

Step 2 is the one nobody mentions

Not every question triggers a search. Some answers come straight from model weights, in which case no source is retrieved and no citation exists to win. This is why some queries in your category will never produce a citation for anyone.

Step 3 is invisible and consequential

The assistant rewrites your question into its own search queries. A user asking a long conversational question may generate several short queries you would never have thought to target.

Steps 4 to 6 are your entire opportunity

Be retrievable, be extractable, look reliable. Everything sold as answer engine optimization is some version of these three.

Step 8 is closed

Attribution happens inside the model. There is no submission process, no ad inventory and no supplier with access to it.

Retrieval-augmented generation, without the jargon

Your question is converted into a numeric representation of its meaning. Passages with similar meaning are retrieved. The strongest handful are packed into a limited context budget. The model then answers from that packet — and only from that packet.

That last clause is the whole game. If your passage did not make the context window, the model did not weigh you and reject you. It never saw you. Your competition is not everyone who has written about the topic; it is whoever’s passage fits the budget.

It also explains why keyword density does nothing here. Similarity is computed on meaning, so repeating a phrase does not increase your odds of retrieval.

Retrieval-augmented generation, without the jargon
This is the single most useful mental model in AEO: your competition is not everyone who wrote about the topic, it is whoever’s passage fits the context budget.
Passage not page — Mental model. Retrieval selects a section.
Meaning not keywords — Mental model. Similarity search, not term frequency.
Context budget — Mental model. Most candidates are discarded before the model reads.
Named or absent — Mental model. No position eight to settle for.
Eligibility first — Mental model. Preference is irrelevant if you were never a candidate.
Corroboration is reliability — Mental model. Third parties are the trust signal.

Why retrieval selects passages, not pages

A model has a limited amount of room. Packing whole pages into it would waste most of that room on text irrelevant to the question, so systems extract the section that appears to answer it.

This is the single most actionable fact in AEO, and it has an uncomfortable consequence: a thinner competitor whose section answers its own heading in the first sentence will beat your better, deeper page whose answer arrives in paragraph four. That is a structural problem, not a quality judgement, and it is fixable in an afternoon.

Why passage structure decides everything
Top-right wins. The uncomfortable implication is that a thinner, better-structured competitor beats your superior page routinely, and that is a fixable problem rather than a fairness problem.
403 to the crawler — Failure. Never becomes a candidate.
JS-only content — Failure. Retrieved, but empty.
Answer in paragraph four — Failure. Extracted badly or not at all.
Answer split across sections — Failure. Nothing clean to lift.
Unsourced claims — Failure. Filtered at the reliability step.
Contradictory entity data — Failure. Model hedges or omits you.
Section answers its heading — Works. Directly serves extraction.
Tables for comparisons — Works. Self-contained, survives summarisation.
Plain definitions — Works. Quotable in one sentence.
Inline sourcing — Works. Survives the reliability filter.
Consistent entity data — Works. Removes the reason to hedge.
Third-party corroboration — Works. The strongest reliability signal available.

How the major surfaces differ mechanically

They are not one system. Perplexity leans heavily on fetching live at question time. Google’s AI surfaces draw on the Google index. Copilot leans on Bing. ChatGPT uses a crawled index plus live retrieval depending on the query.

The practical consequence is that absence on one surface and presence on others is diagnostic rather than random. It usually points at that platform’s crawler or its index, not at your page.

It also explains a pattern we see repeatedly: smaller sites tend to appear in Perplexity first, because a live fetch depends far less on deep indexation than an index-led surface does.

How the major surfaces differ mechanically
Indicative, from documented and observable behaviour rather than published architecture. The practical point: Perplexity’s live fetch is why small sites often see movement there first.
ChatGPT — GPTBot, OAI-SearchBot. Crawled index plus live retrieval.
Perplexity — PerplexityBot. Live fetch at question time.
AI Overviews — Googlebot, Google-Extended. Google index.
Copilot — Bingbot. Bing index plus partner API.
Claude — ClaudeBot. Web search when enabled.
Gemini — Google-Extended. Google grounding.

Live-fetching surfaces

Retrieve at question time. Kinder to new and low-authority pages, because being reachable matters more than being long-established.

Index-led surfaces

Retrieve from a pre-built index. Ordinary SEO indexation is a hard prerequisite, which is why answer engine work sits on top of search work rather than replacing it.

Hybrid surfaces

Use both, and switch depending on the query. This is why results on the same question can differ between sessions on the same platform.

Why this matters for testing

If you test one platform and conclude you are invisible, you have measured one retrieval architecture. Test at least two.

Why your new page has not appeared yet

There is a chain between publishing and being citable, and every link takes time: crawl, index or cache, retrieval eligibility, then selection against everything else. On answers drawn from model weights rather than retrieval, add a model refresh cycle measured in months.

So ‘we published it last week and nothing happened’ is the expected outcome rather than a failure. The useful early signal is not a citation — it is a crawler hit in your server logs, which usually arrives within days.

Why your change did not show up yet
This chain is why ‘we published it last week and nothing happened’ is the wrong expectation rather than a failure.

Days: crawl

Check your logs for the AI user agents. This is the first real evidence that anything is happening.

Weeks: eligibility

The page becomes a retrieval candidate. Still no guarantee of selection.

Weeks to months: selection

It has to beat other candidates on relevance and reliability, which is where structure and corroboration do their work.

Months: weight-based answers

Anything the model answers from memory rather than retrieval changes only on a refresh, and nobody outside the labs schedules those.

What is documented, what is observed, and what is inference

Crawler behaviour and robots.txt handling are documented by the providers. Passage-level retrieval is standard in the published literature. Whether a given surface executes JavaScript is directly observable by testing. The count of sources in an answer you can simply count.

What is not available to anyone outside the labs: the ranking weights inside retrieval, and the reason a specific source was selected over another. We have labelled every claim on this page by which category it falls into, because the alternative is the confident nonsense this field is full of.

What is documented, observed, or inferred
We have labelled each claim on this page by how it is known. Where something is inference we say inference, because the alternative is the confident nonsense this field is full of.
Documented — Crawler behaviour. Published by the providers.
Documented — Passage retrieval. Standard in RAG literature.
Observable — JavaScript execution. Test it yourself with curl.
Observable — Source count per answer. Count them.
Not published — Ranking weights. No provider discloses them.
Not exposed — Selection reasoning. Nobody can tell you why.

Where this leaves the practical work

Three things, in order of how cheaply they are fixed. Be reachable: allow the crawlers and return a 200. Be readable: render your content without JavaScript. Be liftable: make each section answer its own heading in its first sentence.

Then the slower half — look reliable. Source your claims inline, keep your business description identical everywhere, and accumulate third parties saying about you what you say about yourself.

The mechanics, in numbers
The last zero is the honest boundary of this page. Nobody outside the labs knows the weights, so anyone quoting them precisely is guessing.

Reachable

robots.txt permits the five crawlers, and your server and CDN return 200 rather than 403 to them. Free, minutes, binary.

Readable

A scriptless fetch returns your actual content. This is the expensive one and the commonest severe failure.

Liftable

Each section answers the question its heading asks, in the first sentence. Cheap, affects every page, and improves the page for human readers too.

Reliable

Inline sourcing, consistent entity data, and independent corroboration. Slow, durable, and the thing that separates a candidate from a chosen source.

A note on which query types this mechanism favours

Our own September 2026 capture found the answer differs sharply by query type. Definition queries are publisher-locked — the lowest page-one domain authority we measured on ‘what is answer engine optimization’ was 56. Commercial service and tool queries are considerably more open, with domains at 17 and 18 holding page-one positions.

Mechanism and procedural questions of the kind this page answers sit in between. We are stating that because it is the honest framing: this page is written as mechanism rather than definition deliberately, and our published dataset is the reason why.

How open each query type is, from our own September 2026 capture
From our published dataset. Mechanism and procedural queries sit in the middle; definition queries are the locked end. This page is deliberately written as mechanism rather than definition.

Find out where in the pipeline you are actually failing

Steps four, five and six are testable on your own site in an afternoon. An AI visibility audit runs them, reads your server logs, and tells you which link in the chain is broken.

/ai-visibility-audit

Want this done for your site?We build and maintain the search, content and paid programmes described on this page.

Get a free proposal

Everything else we have written on search, AI and getting found

AI, AEO and what is changing

Websites and design

Choosing and working with an agency

Social, content and brand

By industry and by situation

Not sure which of these applies to you?Tell us the situation and we will say plainly what we would do first, and what we would not.

Talk it through

The eight steps, and who controls each one

Eight steps between a question and a citation, with who controls each and what you can actually influence.

The pipeline as a control map
StepWhat happensWho controls itWhat you can do
1. Question arrivesUser types a full sentenceUserNothing
2. Search decisionSystem decides whether to retrieve at allProviderNothing
3. Query generationYour question is rewritten into search queriesProviderNothing
4. RetrievalCandidates pulled from index or live fetchSharedBe crawlable and indexed
5. ExtractionPassages pulled from candidate pagesSharedStructure a clean passage
6. Ranking and filteringPassages scored on relevance and reliabilitySharedSource claims, build corroboration
7. CompositionModel writes prose from surviving passagesProviderNothing
8. AttributionThree or four sources namedProviderNothing. Not for sale

The same question, four different machines

How four systems handle an identical question, and why the answers diverge.

How the major surfaces retrieve
SurfaceCrawlerPrimary retrievalBest early signal for a small site
ChatGPT with searchGPTBot, OAI-SearchBotCrawled index plus live retrievalCrawler hits in logs
PerplexityPerplexityBotLive fetch at question timeOften the first surface to cite you
Google AI OverviewsGooglebot, Google-ExtendedGoogle indexRequires ordinary indexation first
Google AI ModeGooglebot, Google-ExtendedGoogle indexSame as above
CopilotBingbotBing index plus partner APICheck Bing indexation specifically
ClaudeClaudeBotWeb search when enabledClaudeBot must be allowed
GeminiGoogle-ExtendedGoogle groundingGoogle-Extended must be allowed

Failure modes, mapped to the step they break

Each common failure traced to the specific step it interrupts.

Where each common problem actually bites
ProblemBreaks atSymptomFix cost
robots.txt disallowStep 4Absent everywhere on that platformFree
403 to AI user agentStep 4Absent, invisible in a browserLow
JavaScript-only contentStep 5Retrieved but emptyHigh
Answer in paragraph fourStep 5Rarely quoted despite ranking wellLow
Answer split across sectionsStep 5Competitor quoted insteadLow
Unsourced statisticsStep 6Named less often on factual queriesLow
Contradictory entity dataStep 6Described wrongly, or hedgedMedium
No third-party mentionsStep 6Consistently behind better-corroborated rivalsHigh

How we know what we know

What this account is based on, what is documented by vendors, and what is our own observation.

Every claim on this page, by evidence class
ClaimEvidence classHow you can verify it
Crawler names and robots.txt behaviourDocumented by providersRead their published docs
Retrieval selects passages not pagesDocumented in RAG literatureRead the published research
Many AI crawlers do not execute JavaScriptObservableFetch your page with curl
Three or four sources typically namedObservableCount them across answers
Live-fetch surfaces favour reachabilityObserved patternTest a new page across platforms
Ranking weights inside retrievalNot publishedNobody outside the labs knows
Why one source was chosen over anotherNot exposedNobody can tell you, including us

Video: AI search, retrieval and how answers get built

Background viewing. The pipeline is written out in full above; these are context rather than the source of anything here.

Getting found in search

AI, AEO and what is changing

Paid media and lead generation

Websites and design

Choosing and working with an agency

Social, content and brand

By industry and by situation

Frequently asked questions

How does AI search work?
Eight steps: the question arrives, the system decides whether to search, it rewrites your question into its own queries, retrieves candidates from an index or a live fetch, extracts passages rather than whole pages, ranks those passages on relevance and apparent reliability, composes an answer from the survivors, and names three or four sources.
How does ChatGPT choose which sources to cite?
It retrieves candidate passages, packs the strongest into a limited context budget, and answers from that packet — sources named are generally those whose passages survived. The exact ranking weights are not published by any provider, so anyone stating them precisely is guessing.
Why does AI cite a worse page than mine?
Usually structure. Retrieval selects a passage, not a page, so a thinner competitor whose section answers its own heading in the first sentence beats a better page whose answer arrives in paragraph four. It is a structural problem rather than a quality judgement.
What is retrieval-augmented generation?
The pattern behind most cited answers. Your question becomes a numeric representation of meaning, similar passages are retrieved, the strongest are packed into a limited context budget, and the model answers from that packet rather than from memory.
Does keyword density affect AI search?
No. Retrieval matches on meaning rather than term frequency, so repeating a phrase does not increase your odds of being retrieved.
Why does a passage matter more than a page?
Because context space is limited. Packing whole pages would waste most of that space on text irrelevant to the question, so systems extract the section that appears to answer it.
Do all AI assistants work the same way?
No. Perplexity leans on live fetching at question time, Google’s AI surfaces draw on the Google index, Copilot leans on Bing, and ChatGPT uses a crawled index plus live retrieval depending on the query.
Which platform is easiest for a small site to appear in?
Usually Perplexity, because live fetching depends far less on deep indexation than an index-led surface does.
Why did my new page not get cited immediately?
There is a chain: crawl, then index or cache, then retrieval eligibility, then selection against every other candidate. Days for the crawl, weeks for eligibility, longer for selection. The first real signal is a crawler hit in your server logs.
Can I influence which sources an AI names?
Only indirectly, at three steps: being retrievable, being extractable, and appearing reliable. The final attribution step happens inside the model and is not addressable by anyone.
Can I pay to be cited?
No. There is no ad inventory, no submission process and no paid tier at the attribution step.
What is the context budget?
The limited amount of retrieved text a model reads before answering. Most candidate passages are discarded before this point, which means your real competition is whoever’s passage fits — not everyone who wrote about the topic.
Does every question trigger a search?
No. Some answers come from model weights alone, with no retrieval and therefore no citation for anyone. This is why certain queries in your category will never produce a citation.
What are the queries the assistant actually searches for?
Rewritten versions of your question, generated by the system and not shown to you. A long conversational question may produce several short queries you would not have targeted.
How do I know if AI crawlers reach my site?
Grep your server logs for GPTBot, OAI-SearchBot, ClaudeBot, PerplexityBot and Google-Extended over the last thirty days. That is the only direct evidence.
Does JavaScript break AI search?
Frequently, yes. Many AI crawlers do not execute it, so a page that assembles content in the browser returns almost nothing to them while looking perfect to you.
How many sources does an AI answer name?
Typically three or four. There is no second page, so there is no long tail of positions to settle for.
Is any of this published by the AI companies?
Partly. Crawler names and robots.txt behaviour are documented. Passage-level retrieval is standard in the published literature. Ranking weights and selection reasoning are not published by anyone.
Why do answers change between runs of the same question?
Retrieval is influenced by phrasing, timing, region, account state and model version. Identical prompts genuinely return different sources, which is why single checks prove very little.
What is the highest-leverage change I can make?
Allow the crawlers, then make each section answer its own heading in its first sentence. The first is free and binary; the second is cheap and affects every page you own.
Does site speed matter for AI search?
Indirectly. It affects crawling and indexation, which affect eligibility. It is not a selection factor in the way structure and corroboration are.
Why does corroboration matter at the ranking step?
Because reliability is part of what passages are filtered on, and independent third parties repeating what you say is the strongest reliability signal available to a system that cannot verify claims directly.

Sources and further reading

  1. Google Search Essentials — SEO starter guide
  2. Google: creating helpful, reliable, people-first content
  3. Google: intro to structured data
  4. Google: LocalBusiness structured data
  5. Google: FAQPage structured data
  6. Google: Article structured data
  7. Google: Product structured data
  8. Google: title links in search results
  9. Google: control your snippets
  10. Google: robots.txt introduction
  11. Google: sitemaps overview
  12. Google: consolidate duplicate URLs
  13. Google: redirects and Search
  14. Google: JavaScript SEO basics
  15. Google: multi-regional and multilingual sites
  16. Google Search Central Blog
  17. Google: get started with Search Console
  18. Google: how local search results are determined
  19. Google Business Profile: prohibited and restricted content
  20. Google Business Profile: address and service area guidelines
  21. Google Business Profile: review policy
  22. Google Business Profile: add or edit categories
  23. Google Ads: location targeting settings
  24. Google Ads: about negative keywords
  25. Google Ads: about Quality Score
  26. Google Ads: importing offline conversions
  27. Google Ads: about Smart Bidding
  28. Google Ads: about Performance Max
  29. Google Local Services Ads: eligibility and screening
  30. Google Ads: keyword match types
  31. Google Analytics 4: about conversions
  32. Google Analytics 4: attribution models
  33. US Census Bureau QuickFacts: New Jersey
  34. US Census Bureau: American Community Survey
  35. US Census: Statistics of US Businesses
  36. Bureau of Labor Statistics: New Jersey data
  37. BLS: Occupational Employment and Wage Statistics
  38. NJ Department of Labor: labor market information
  39. New Jersey Business Action Center
  40. US Small Business Administration: New Jersey district
  41. USA.gov: business resources
  42. web.dev: Core Web Vitals explained
  43. web.dev: Largest Contentful Paint
  44. web.dev: Cumulative Layout Shift
  45. web.dev: Interaction to Next Paint
  46. Google PageSpeed Insights
  47. Google Rich Results Test
  48. Google Search Console
  49. W3C Markup Validation Service
  50. Schema.org: LocalBusiness type
  51. Schema.org: Service type
  52. Schema.org: FAQPage type
  53. Schema.org: HowTo type
  54. W3C: WCAG 2.2 quick reference
  55. FTC: CAN-SPAM Act compliance guide
  56. FCC: telemarketing and robocall rules (TCPA)
  57. FTC endorsement guides — reviews and testimonials
  58. FTC: rule on consumer reviews and testimonials
  59. HHS: HIPAA guidance on online tracking technologies
  60. New Jersey Courts: attorney advertising guidelines
  61. New Jersey DCA: construction codes and permits
  62. New Jersey Home Improvement Contractor registration
  63. New Jersey Division of Consumer Affairs
  64. TikTok for Business
  65. TikTok Creative Center
  66. TikTok Ads Help Center
  67. TikTok Community Guidelines
  68. TikTok Terms of Service
  69. TikTok Privacy Policy
  70. TikTok Safety Center
  71. TikTok Transparency Center
  72. TikTok Creator Portal
  73. TikTok Newsroom
  74. TikTok for Developers
  75. TikTok advertising solutions
  76. TikTok Creator Marketplace
  77. TikTok Business Center
  78. TikTok for Business blog
  79. TikTok Creative Center: top ads
  80. TikTok Branded Content policy
  81. TikTok Shop for sellers
  82. Instagram for Business
  83. Instagram for Creators
  84. Instagram Help Center
  85. About Instagram
  86. Meta Business Suite
  87. Meta Business Help Center
  88. Meta Transparency Center
  89. About Meta
  90. Meta: Instagram platform docs
  91. YouTube Creators
  92. YouTube Official Blog
  93. YouTube Shorts help
  94. How YouTube Works
  95. YouTube Studio
  96. LinkedIn Marketing Solutions
  97. LinkedIn Help
  98. Pinterest Business
  99. Pinterest Business Help
  100. Snapchat for Business
  101. X for Business
  102. Reddit communities
  103. Reddit for Business Help
  104. ASCAP
  105. BMI
  106. SESAC
  107. Global Music Rights
  108. PRS for Music (UK)
  109. PPL (UK)
  110. SOCAN (Canada)
  111. APRA AMCOS (Australia)
  112. GEMA (Germany)
  113. SACEM (France)
  114. SIAE (Italy)
  115. JASRAC (Japan)
  116. IFPI
  117. RIAA
  118. National Music Publishers Association
  119. Harry Fox Agency
  120. SoundExchange
  121. Music Reports
  122. Epidemic Sound
  123. Artlist
  124. Soundstripe
  125. PremiumBeat
  126. AudioJungle
  127. Free Music Archive
  128. Creative Commons
  129. Incompetech
  130. FTC: advertising and marketing
  131. FTC: disclosures 101
  132. FTC: endorsement guides
  133. FTC: consumer reviews rule
  134. FTC: advertising FAQs
  135. US Copyright Office
  136. US Copyright Office: DMCA
  137. US Copyright Office: music FAQ
  138. US Copyright Office: fair use FAQ
  139. USPTO: trademarks
  140. UK Advertising Standards Authority
  141. ACCC (Australia)
  142. Competition Bureau Canada
  143. GDPR overview
  144. California Consumer Privacy Act
  145. COPPA
  146. FTC: children’s privacy
  147. W3C Web Accessibility Initiative
  148. W3C: WCAG
  149. W3C: captions
  150. W3C: making audio and video accessible
  151. ADA.gov
  152. WebAIM
  153. Epilepsy Foundation
  154. Pew Research: internet and technology
  155. DataReportal
  156. US Census Bureau
  157. US Bureau of Labor Statistics
  158. Interactive Advertising Bureau
  159. Think with Google
  160. Google Trends
  161. Nielsen insights
  162. Schema.org: VideoObject
  163. Schema.org: SocialMediaPosting
  164. Schema.org: MusicRecording
  165. Schema.org: HowTo
  166. Schema.org: FAQPage
  167. Schema.org: Organization
  168. Google: video best practices
  169. Google: video structured data
  170. CapCut
  171. Adobe Premiere Rush
  172. DaVinci Resolve
  173. Canva
  174. Descript
  175. VEED
  176. Kapwing
  177. Otter.ai
  178. Later
  179. Buffer
  180. Hootsuite
  181. Sprout Social
  182. Google Analytics
  183. Google Search Console
  184. Google Analytics developer docs
  185. GA4: events and conversions
  186. Matomo
  187. Plausible Analytics
  188. Similarweb
  189. UK Information Commissioner’s Office
  190. Office of the Privacy Commissioner of Canada
  191. Australian OAIC
  192. European Data Protection Board
  193. EU data protection
  194. EU Digital Services Act
  195. Ofcom
  196. FCC
  197. AIGA
  198. Nielsen Norman Group
  199. Smashing Magazine
  200. web.dev
  201. MDN: web media
  202. MDN: the video element
  203. ISO 21001 (reference)
  204. Buma/Stemra (Netherlands)
  205. STIM (Sweden)
  206. Teosto (Finland)
  207. Koda (Denmark)
  208. TONO (Norway)
  209. IMRO (Ireland)
  210. SGAE (Spain)
  211. ZAiKS (Poland)
  212. KOMCA (South Korea)
  213. MCSC (China)
  214. CISAC
  215. World Intellectual Property Organization
  216. TikTok: creating videos
  217. TikTok: exploring videos
  218. TikTok: privacy settings
  219. TikTok: growing your audience
  220. TikTok Creator Academy
  221. TikTok Effect House
  222. TikTok for small business
  223. Instagram: Reels help
  224. YouTube: Shorts best practice
  225. How YouTube recommends
  226. Pinterest Predicts
  227. Snapchat for Business
  228. Hootsuite blog
  229. Social Media Examiner
  230. Marketing Week
  231. Adweek
  232. Google Search Essentials — SEO starter guide
  233. Google: creating helpful, reliable, people-first content
  234. Google: intro to structured data
  235. Google: LocalBusiness structured data
  236. Google: FAQPage structured data
  237. Google: Article structured data
  238. Google: Product structured data
  239. Google: title links in search results
  240. Google: control your snippets
  241. Google: robots.txt introduction
  242. Google: sitemaps overview
  243. Google: consolidate duplicate URLs
  244. Google: redirects and Search
  245. Google: JavaScript SEO basics
  246. Google: multi-regional and multilingual sites
  247. Google Search Central Blog
  248. Google: get started with Search Console
  249. Google: how local search results are determined
  250. Google Business Profile: prohibited and restricted content
  251. Google Business Profile: address and service area guidelines
  252. Google Business Profile: review policy
  253. Google Business Profile: add or edit categories
  254. web.dev: Core Web Vitals explained
  255. web.dev: Largest Contentful Paint
  256. web.dev: Cumulative Layout Shift
  257. web.dev: Interaction to Next Paint
  258. Google PageSpeed Insights
  259. Google Rich Results Test
  260. Google Search Console
  261. W3C Markup Validation Service
  262. Schema.org: LocalBusiness type
  263. Schema.org: Service type
  264. Schema.org: FAQPage type
  265. Schema.org: HowTo type
  266. W3C: WCAG 2.2 quick reference
  267. US Census Bureau QuickFacts: New Jersey
  268. US Census Bureau: American Community Survey
  269. US Census: Statistics of US Businesses
  270. Bureau of Labor Statistics: New Jersey data
  271. BLS: Occupational Employment and Wage Statistics
  272. NJ Department of Labor: labor market information
  273. New Jersey Business Action Center
  274. US Small Business Administration: New Jersey district
  275. USA.gov: business resources
  276. OpenAI — GPTBot and OAI-SearchBot documentation
  277. Google — crawlers and user agents
  278. Retrieval-Augmented Generation (original paper)
  279. GEO: Generative Engine Optimization (research paper)

Want this done for your site?We build and maintain the search, content and paid programmes described on this page.

Get a free proposal

Get a free marketing proposal

Tell us what you are trying to grow and we will come back with a plan, not a pitch deck. Same-day reply on weekdays.

Our website uses cookies from third party services to improve your browsing experience. Read more about this and how you can control cookies by clicking "Privacy Preferences".
Privacy Preferences
When you visit our website, it may store information through your browser from specific services, usually in form of cookies. Here you can change your privacy preferences. Please note that blocking some types of cookies may impact your experience on our website and the services we offer.
Privacy Policy
You have read and agreed to our privacy policy
Required
Contact Us
0