Arvind Kumar

Contact now

Search, answers, citations

Rank. Get quoted. Get picked.

Five acronyms, one job: make sure that when someone asks a question your company can answer, the answer comes from your site, whether the asker is a person on Google or a model composing a reply. This page is how I do that, and this site is the demo. Flip the view toggle in the header to see what a crawler gets.

One loop

All the O’s

SEO, AEO, AIO, GEO and CRO usually get sold as five services by five vendors. They are one system: each hands the visitor to the next, and a gap anywhere leaks the whole funnel.

  1. SEO: Search engine optimization

    Be found.

    Google and Bing can crawl the page, understand it and rank it: fast templates, canonicals that point home, sitemaps that tell the truth. Nothing further down this list works if this layer leaks.

  2. AEO: Answer engine optimization

    Be the answer.

    Each page answers one question in its first screen. Key takeaways and quotable stats sit at the top, and the FAQs are written for the way people ask, so an answer box or an AI Overview can lift the passage whole.

  3. AIO: AI optimization

    Be readable by machines.

    The site explains itself to models and agents: structured data that agrees with the page, an llms.txt that maps it, and forms an AI agent can fill in through WebMCP.

  4. GEO: Generative engine optimization

    Be cited.

    ChatGPT, Gemini, Claude and Perplexity name you as the source. I measure it as citation share against competitors, on a fixed set of real buyer questions, and fill the gaps with pages built to be quoted.

  5. CRO: Conversion rate optimization

    Be chosen.

    All of that work sends a visitor. This layer turns the visit into a lead: calls to action tested against each other, and forms that hold every submission until the CRM confirms it, so no lead goes missing.

  6. One loop, one job

    When a buyer asks a question your company can answer, the answer, the citation and the click all lead back to you.

    Most companies split this between an SEO agency, a content team, a developer queue and a CRO consultant, and every hand-off leaks. I run it as one loop, so a fix in one place never breaks another.

    20+ content hubs built to be cited in AI answers. The technical side of the first three letters is next.

Three letters each, very different jobs

SEO

Be findable

The classic discipline, still the foundation. Crawlable templates, fast pages, canonicals that point home, redirects that survive a migration, sitemaps that tell the truth. Most “AI visibility” problems turn out to be technical SEO problems wearing a new hat.

AEO

Be the answer

Answer engines pull a passage, not a page. So each page answers one question in its first screen, in plain language, with structured data that agrees with the text. Content hubs organised by buyer intent, not by product name.

GEO

Be the source

Generative engines cite. The goal is citation share: of the answers a model gives about your category, how many point at you instead of a competitor. I measure that across models, find the gaps, and fill them with pages built to be quoted.

The operating loop

  1. Measure. Track which questions get answered with us, with a competitor, or with nobody, across the major models and classic search. Weekly, in a sheet, with history.
  2. Find the vacuum. Questions where no vendor is cited are the cheapest wins. Questions where a competitor is cited and we are not are the most urgent.
  3. Build the page that deserves the citation. One intent, a direct answer up top, supporting depth below, schema that matches, internal links from the hub.
  4. Earn the links. Free engineering tools and reference content that other sites want to point at, ungated.
  5. Fix the plumbing. Migrations, canonicals, redirects, consent-safe analytics, sitemaps, llms.txt. Boring, decisive.
  6. Repeat. Re-measure, publish what moved, keep the loop honest.

This site, explained

  • Person and WebSite schema in the head, shown live in the homepage hero
  • Every project page carries its own CreativeWork, WebSite or SoftwareSourceCode schema
  • /llms.txt generated from the live content, not hand-maintained
  • The crawler view in the header: headings labelled, images as alt text, links with their targets
  • Self-hosted fonts, deferred scripts, no page builder, no plugins: see the footer for live page weight

Every week, current role

Semrush AI visibility data goes in, and four things come out:

  • A metrics workbook
  • A week-over-week change report
  • A ranked content plan
  • A short deck for my manager

Once a month, AI Visibility Tracker asks ChatGPT, Gemini, Claude and Perplexity the same buyer questions, so the trend is measured, not guessed.

Search work

6 questions about Arvind Kumar, answered

What is the difference between SEO, AEO and GEO?

SEO makes a site findable: crawlable templates, fast pages, canonicals, redirects and sitemaps. AEO makes each page the answer: one question answered in the first screen, in plain language, with structured data that agrees with the text. GEO makes the site the source AI engines cite, measured as citation share across models.

How do you measure visibility in AI answers?

I track which questions get answered with you, with a competitor or with nobody, across the major AI engines and classic search, with history. A fixed set of real buyer questions, asked of ChatGPT, Gemini, Claude and Perplexity the same way every month, is what makes the trend real.

What gets a page cited by AI engines?

A direct answer up top, supporting depth below, structured data that matches the text, internal links from a hub, and content organised by buyer intent instead of product names. Questions where no vendor is cited yet are the cheapest wins.

Does technical SEO still matter for AI search?

Yes. Most AI visibility problems turn out to be technical SEO problems: crawl access, canonicals, redirects that survive a migration, sitemaps that tell the truth and fast pages.

What is llms.txt?

A plain-text map of a site for language models. This site generates its /llms.txt from the live content instead of keeping it by hand.

What SEO, AEO and GEO results has Arvind had?

In my current role: more than 20 content hubs built around buyer intents, key takeaways and quotable stats at the top of pages, analyst reports published as pages, FAQs restructured for answer engines, and an llms.txt. Before that, organic traffic doubled for a regulated UK fund manager.