Aabhas Vijay

Aabhas Vijay says the keyword is a useful organizing unit for traditional SEO but a misleading starting point for answer engine optimization: AI systems do not research topics the way keyword tools frame them.

• A page aimed at one phrase can miss the surrounding material a model needs to assemble an answer.
• The proposed path is decision, then questions, research queries, evidence and content—the reverse of the familiar keyword-to-article route.
• Vijay draws out two consequences: keyword targeting tends to produce pages that exist because a tool suggested them, and strong keyword rankings do not carry over into AI visibility, since the same prompt can return different results depending on the model, context and retrieval process.

A page built for the keyword "best CRM for startups" may perfectly target that phrase.

• But when someone asks ChatGPT, "Which CRM should I use?", the model may look for best CRM for SaaS, HubSpot alternatives and CRM reviews.
• The page may not cover half of what it needs.

Framed as a correction, not a rejection: Vijay still calls the keyword a useful unit for traditional SEO. The point is where planning starts, not whether to discard keyword research.

Read original

The point is to map a niche's retrieval ecosystem, not just check whether a brand appears.

• If a category never triggers web search, rewriting pages, adding FAQs and changing headings will not earn citations; the answers are coming from training data.
• Vijay also expects a single buyer question to fan out into several searches behind the scenes, treating the resulting queries as multiple opportunities rather than one.

Vijay cites an analysis by Jonathan Berthold of 22,881 AI citations: 34.5 per cent went to domains with a domain authority below 40, and 53.9 per cent of unique cited domains were below DA 40.

• Vijay uses the figures to argue that a small, specific industry publication can be more useful to an AI answer than another generic high-authority site.
• The study is Berthold's work, as reported by Vijay.

Read original

Clarity now splits citations into branded and non-branded buckets: branded means the AI looked up the company by name, while non-branded means it found the company while researching a broader topic.

• The two measure different things: brand pull and category authority.
• A citation counts as branded even when the user never typed the brand name, because the AI still searched for specific product names behind the scenes.
• The update also added Share of Authority, the percentage of citations a domain earns against competitors in queries where it already appears.
• Vijay's directives follow: keep the two citation types separate, review by topic and page rather than by totals, and treat citations, traffic and conversions as three distinct stages, since conversions still live in the CRM.

On a non-branded prompt such as "tools to understand website visitor behavior", AI may still search for specific product names behind the scenes. Clarity now counts that as a branded citation, even though the user never named a brand.

Marks a shift from visibility to business goals: Vijay says the update matters less for new data than for separating signals and connecting them to what actually converts.

Read original

AI collapses discovery and consideration into one conversation, then drops a high-intent buyer onto a product page.

• That page needs to lead with specifications, comparisons and trust signals for visitors who have already done part of the deciding.
• Vijay recommends mining real shopper language from AI queries, Search Console and on-site search; converting it into product attributes such as use case, compatibility, fit, materials and best-for scenarios; rebuilding high-intent product pages; collecting and marking up reviews; and measuring AI and organic side by side by conversion and average order value.
• The same enrichment serves both channels, so the choice is not AI or organic.

The post reports that AI-referred sessions to Shopify storefronts grew 197 per cent year over year, orders grew roughly threefold, and AI-referred shoppers converted about 80 per cent better on product pages than organic-search visitors—closer to double in specification-led categories where buyers compare specs and tradeoffs. It adds that conversion doubled again when AI drew on clean, structured product data instead of scraped feeds.

Read original

The requirement to manually fact-check and review AI-generated material before publication now extends to metadata—including title tags, meta descriptions, structured data and image alt text—not just the visible body.

• Vijay links the change to the September 2026 spam update and its emphasis on human effort, originality and people-first quality.
• The post recommends validating markup against Google's guidelines, noting that Google Merchant Center requires specific IPTC metadata for ecommerce, and adding experiences data to pages.
• Review and validation belong between drafting and publishing, and the same rules apply to parts of a page readers do not see.

Vijay reports that pages were already being deindexed or blacklisted after the spam update and recommends adding experiential data to sites as the next step.

Read original

This post argues that keywords are a useful organizing unit for traditional SEO but a misleading starting point for answer engine optimization, because AI systems do not research a topic the way a keyword tool describes it.

• A page aimed at “best CRM for startups,” for example, can miss what ChatGPT looks for when a user asks “Which CRM should I use?”
• The model may research best CRM for SaaS, HubSpot alternatives, and CRM reviews instead.
• Aabhas Vijay contrasts two production paths—keyword to article versus decision to questions to research queries to evidence to content—and lists six specific failure modes, including creating pages nobody needs, ignoring the sources AI trusts, and mistaking keyword ranking for AI visibility.

For brands planning a campaign

A useful workshop module: rewrite content briefs around buyer decisions and retrieval queries rather than keyword lists, with a concrete before-and-after exercise for a SaaS team.

Aabhas Vijay describes a consultation routine: pick five to ten real buyer questions, run them in ChatGPT, and check whether it triggers a live web search and what it cites afterwards—comparison pages, industry publications, competitors, or brand sites.

• The post claims ChatGPT averages 7.6 fan-out searches per prompt, and cites a Jonathan Berthold analysis of 22,881 AI citations in which 34.5% went to domains with domain authority below 40, while 53.9% of unique cited domains were below DA 40.
• If no searches fire in a category, rewriting pages and adding FAQs will not earn citations; the first job is to establish that a retrieval opportunity exists.

For brands planning a campaign

Gives a sales or content team a short pre-engagement audit and an argument that small, specific publications can out-cite larger sites in AI answers.

This post explains that Microsoft Clarity's August 3 update splits AI citations into branded and non-branded buckets and adds a Share of Authority metric: the percentage of citations a domain earns versus competitors inside queries where it already appears.

• Aabhas Vijay notes that a citation counts as branded even when the user never typed the brand name.
• The post advises readers not to combine the two signals, to review by topic and page rather than totals, and to treat citations, traffic, and conversions as three separate stages.
• Treat Clarity as a sample and ask what other sources validate it.

For brands planning a campaign

Provides a ready-made reporting structure for demonstrating AEO progress to a marketing lead without conflating awareness and conversion metrics.

Aabhas Vijay summarizes Google's revised documentation dated October 1, 2026, which stresses manually fact-checking and reviewing AI-generated content before it goes live.

• The post links that change to the September 2026 spam update and notes that the requirement extends to metadata such as title tags, meta descriptions, structured data, and image alt text, not just the main body.
• It also flags that Google Merchant Centre requires specific IPTC metadata for ecommerce, reports pages already being deindexed or blacklisted after the spam update, and recommends adding experiential data to websites.

For brands planning a campaign

A governance checklist for content teams moving AI-assisted drafts into production, and a natural closing module for a workshop on doing AEO without risking search visibility.

The post uses Shopify's Q2 data to argue that a small AI-referral share can still carry disproportionate revenue: AI-referred sessions to Shopify storefronts grew 197% year over year, orders grew threefold, and AI-referred shoppers convert about 80% better on product pages than organic-search visitors—roughly double in specification-led categories where buyers compare specs before deciding.

• Aabhas Vijay notes that conversion doubled again when AI pulled from clean structured product data instead of scraped feeds.
• The post closes with five actions: map real shopper language from AI queries, Search Console, and site search; convert it into product attributes; rebuild product pages for high-intent arrivals; collect and mark up reviews; and measure AI and organic side by side by conversion and average order value.

For brands planning a campaign

Gives a practical structure for a business case, with concrete data sources and page-level changes, and shows how the same catalogue work serves both AI and organic discovery.

The post contrasts a keyword-to-article path with a sequence of decision, questions, research queries, evidence, and content.

• It gives the example of a page built for "best CRM for startups" that misses the surrounding material an AI assembles an answer from.
• They also separate keyword ranking from AI visibility, noting that the same prompt can produce different results depending on model, context, and retrieval process.

Map the decision first, derive the questions and evidence behind it, then build content around retrieval queries rather than a seed phrase.

For brands planning a campaign

A workshop on this framing would give a marketing team a usable starting point beyond a keyword list: a decision map and research queries to test live.

Examples: Source 1

They cite Jonathan Berthold's analysis of 22,881 AI citations, in which 34.5% went to domains with DA below 40, to argue that relevance and specificity can outweigh raw authority.

Buyer questions, observed search behavior, fan-out mapping, a retrieval check, then a cited-source inventory.

If a category never triggers web retrieval, Vijay says rewriting pages, adding FAQs, and changing headings will not produce citations. The retrieval check is a precondition, not a later optimization.

For brands planning a campaign

This is the most portable piece of the material: it gives a team a low-cost diagnostic to run before committing budget to AEO content and works as a first hands-on exercise in a session.

Examples: Source 1

They describe Share of Authority as the share of citations a domain earns against competitors within the queries where it already appears. Their guidance is to keep the two citation types apart, review by topic and page rather than totals, treat Clarity as a sample, and track citations, traffic, and conversions as three separate stages, with conversions still living in the CRM.

Segment metrics by query type and funnel stage, and review them at topic and page level instead of relying on aggregate totals.

For brands planning a campaign

A measurement module built on this would give a B2B SaaS team a defensible reporting frame instead of a single citation count, a common gap when teams start reporting on AI visibility.

Examples: Source 1

They note conversion doubled again when AI drew on clean, structured product data rather than scraped feeds, and give a five-step action list covering query mining, attribute structuring, product-page rebuilds for high-intent arrivals, review collection and markup, and side-by-side measurement of AI and organic.

Pull AI queries from Shopify's agentic admin, Search Console, and on-site search; convert them into attributes such as use case, compatibility, materials, and best-for scenarios; then rebuild product pages around specs, comparisons, and trust signals.

For brands planning a campaign

Strongest for a merchant or commerce team. For a pure B2B SaaS brief, the transferable elements are attribute structuring and the insistence on measuring AI and organic together rather than choosing between them.

Examples: Source 1

They link it to the September 2026 spam update's emphasis on human effort and originality, note that Google Merchant Center requires specific IPTC metadata for ecommerce, and report pages already being deindexed.

Add a review and validation step after any AI drafting, and check markup against Google's stated requirements before publishing.

For brands planning a campaign

Adds a governance checkpoint to any AI-assisted content program: the session should cover review, fact-checking, and markup validation, not just generation speed.

Examples: Source 1

The numbered sets include five ways keyword thinking harms AEO, five steps to test whether ChatGPT searches a category, five actions after the Shopify data, and five points on Google's revised guidance. Several respond to dated platform changes, including Microsoft Clarity's August 3 update, Google's October 1 documentation revision, and Shopify's Q2 data, so the material tracks industry movement rather than presenting one fixed framework.

These posts do not state their individual publication dates, so any development over time is visible only through which platform updates each post answers. That is a weaker basis than a dated chronology.

For brands planning a campaign

A talk built around this format is easy to run as a working session: the numbered steps give a team a framework to apply to its own categories during the hour.

Examples: Source 1 · Source 2 · Source 3 · Source 4 · Source 5

They also argue that AI search results must be measured through several signals at once, not a single ranking or citation count.

• Cohen's measurement post names AI-referred traffic, branded and direct traffic, self-reported attribution, and manual spot checks as signals to triangulate.
• That echoes Vijay's insistence on tracking citations, traffic and conversions as separate stages.

Cohen stays closer to keyword selection and portfolio choices — which terms a small site can realistically rank for, comparison pages, and phrases mined from customer conversations.

• Vijay focuses more on whether AI retrieval fires at all, how to structure product and page data so a model can read it, and how to respond to specific platform changes.
• Cohen comes across as an operator improving targeting; Vijay, as a diagnostician testing how the retrieval system behaves.

Read original · Original 2

They connect AI visibility to entity clarity, consistent category language, and product data detailed enough to answer what, who and why. Both use ecommerce examples to make the point.

Mohsin Nisar M. packages the material as reusable frameworks and checklists, including a full product-page optimization framework covering entity-attribute-value structure, intent mapping, schema and reviews, and offers a downloadable version. Vijay's posts are shorter diagnostics tied to a recent platform or data release, with less emphasis on a permanent checklist.

Read original · Original 2

Existing authority is often wasted because the information layer is unclear, not because the site lacks backlinks. They connect AI visibility to ordinary SEO fundamentals — crawlable pages, accurate entity information, and sensible keyword and intent choices.

Azhan Shafique writes mostly as an agency founder explaining client engagements, including decisions about when AI tools help and when human judgment has to set strategy.

• They also cover classic SEO work such as keyword difficulty, intent checks and backlink selection.
• Vijay writes shorter, more structured posts, with a stronger focus on measurement design and interpreting what platform updates mean for reporting.

Read original · Original 2

One post walks through a founder who spent six months on "CRM software" with about 25 backlinks against a top result with more than a million, then shows the alternative: niche buyer-intent terms such as "CRM for solar installers" at far lower difficulty, competitor-comparison searches ("[Competitor] alternatives," "[Tool A] vs [Tool B]"), and exact phrases mined from customer conversations.

• The second post answers the measurement question a SaaS marketing lead is likely to ask next, triangulating AI-referred traffic (citation clicks carry utm_source, trackable in analytics), lifts in direct and branded traffic from people who saw the brand in a model and searched later, a self-reported "how did you hear about us?" field, and periodic manual spot checks.
• That combination suits a team seeking quick revenue-linked targets and a lightweight way to show whether AI search effort is working, without commissioning a workshop.

The tradeoff is depth and structure.

• Cohen's material offers tactical, post-length guidance, not a runnable session: it has no five-step protocol for testing whether a category triggers retrieval, no query fan-out analysis, and no reporting frame that separates branded from non-branded citations or measures competitive share within queries where the brand already appears.
• The measurement advice is triangulation heuristics—what to look at—rather than a sheet the team can populate.
• A SaaS team seeking decision maps, a prompt-and-citation log, and named publishing rules will get more from a workshop in Vijay's style; a team that mainly needs lower-difficulty keyword targets and a first-pass ROI story can act on Cohen's posts directly.

Examples: Source 1 · Source 2

The stronger post reframes the question from "how do I optimize for ChatGPT?" to "why would an AI recommend me over my competitors?" and argues that the differentiator is solution specificity: not "best coffee machine" but "best coffee machines under $300 with a grinder that fits a small kitchen," with Amazon's product data, Reddit threads, and consistent category terminology cited as what actually surfaces.

• The follow-on post is a product-page optimization framework for the move from strategy to pages: entity-attribute-value structure, query-intent mapping across transactional, commercial, informational and navigational searches, schema markup, image optimization, reviews and user-generated content, technical checklists, product-variant handling without duplicate content, and a discontinued-product strategy to preserve link equity, all explicitly including AI search and conversational commerce readiness.
• For a SaaS team with technically thin pages or inconsistent category language, this maps more directly to the fix than a general AEO diagnostic.

The tradeoff is that this is a framework and an argument, not a tested workshop, and it requires a conversion step: the detailed version is offered as a PDF distributed after connecting and commenting.

• That makes it hard to evaluate the substance beforehand; there is no live exercise, prompt-testing protocol, or measurement frame comparable to the one described in the fit section.
• The positioning case is persuasive strategic framing, not a reproducible test the team can run on its own category.
• Its supporting evidence comes from cited third-party analysis rather than the creator's documented client work in the posts.
• Ask for a sample of the PDF before connecting and commenting.

Examples: Source 1 · Source 2

One post is a diagnostic case: a surgical instrument manufacturer with strong backlinks, ISO certifications and regulatory references that ChatGPT and Google AI still misunderstood.

• The problem was resolved not by building more links but by fixing the information layer—clear statements of who the company is, what it manufactures, its location, product categories, certifications, expertise and history—then re-indexing important pages and seeing accurate identification appear in AI answers.
• The conclusion, that the goal is entity understanding rather than mentions, applies directly to a SaaS brand that AI answers confuse or ignore.
• The second post addresses a real internal failure mode: a prospect who thought Claude replaced an SEO agency and received high-volume keywords with no difficulty or intent data.
• Shafique argues for low-competition terms, intent checked one by one, and topically organized pages, with AI as speed and human judgment as the strategy layer.
• That framing can help convince a skeptical internal stakeholder.

The tradeoff is a narrow evidence base.

• The posts describe two accounts of the same problem area, built on a single client case rather than a repeatable method.
• The reasoning is sound, but the proof of general applicability is thinner than for the other alternatives.
• There is no measurement framework, reporting sheet, publishing checklist, or session structure, so a buyer seeking a hands-on workshop with deliverables should ask for those.
• This is also the closest alternative in substance to the primary candidate: it addresses specific symptoms (entity confusion or over-reliance on AI-generated keyword lists) rather than the entire need.

Examples: Source 1 · Source 2