Technical SEO and AEO Checklist for Indian Business Websites (2026)
A practical checklist covering indexability, structured data, Core Web Vitals and Answer Engine Optimisation - making your site legible to both Google and AI answer engines.
What this covers
- What changed: AI answers sit in front of search
- The technical foundation that still decides everything
- Structured data: how machines read your business
- Core Web Vitals in practice
- Being citable by AI answer engines
- The full checklist
Technical SEO used to mean making a site legible to one crawler. Now a growing share of the people looking for your business never see a results page at all — they ask an assistant, get a paragraph, and either see your name in it or they do not.
The good news is that the work overlaps heavily. A site built so Google can parse it confidently is also a site an AI system can summarise accurately. Here is the checklist we run on every Indian business site we audit.
1. Make sure you can be crawled and indexed
It sounds basic. It is also the single most common serious problem we find, including on sites that have been live for years.
- robots.txt exists and is not blocking anything important. A missing robots.txt returns a 404 and gives crawlers no sitemap pointer.
- An XML sitemap exists, lists every page you want indexed, and is submitted in Google Search Console.
- Every page has a canonical URL. Without one, the same content reachable at
/and/home.phpis treated as two competing pages, splitting your ranking signals. - One hostname. Pick www or non-www and 301 the other to it. Both resolving is duplicate content.
- Search Console is verified and the Coverage report has been read. It will tell you exactly which pages Google refuses to index and why.
2. Give every page its own title and description
We routinely audit sites where every page shares one title tag. When your homepage, services page and seven service detail pages all say the same thing, none of them targets anything a person would search, and Google treats the set as near-duplicates.
- Unique title per page, roughly 50–60 characters, with the primary keyword near the front
- Unique meta description, roughly 150–160 characters — it does not rank you directly, but it decides whether people click
- Exactly one H1 per page, describing what the page is actually about
- A sensible heading hierarchy. H2s for sections, H3s beneath them. Do not use heading tags for styling — a big number in an H2 tells a crawler that number is a section topic
3. Structured data: how machines read your business
Schema markup is the difference between a machine guessing what your business is and being told. For AI answer engines this matters more than it ever did for blue links, because a system generating a one-paragraph answer needs facts it can trust.
At minimum, implement in JSON-LD:
- Organization — legal name, founding date, logo, address, contact, and
sameAslinks to your verified social profiles - LocalBusiness if you serve a geographic area — with full postal address and coordinates
- WebSite and WebPage for site-level context
- BreadcrumbList on every page below the homepage
- Service or OfferCatalog for what you sell
- FAQPage where you have genuine Q&A content
- Article on blog posts, with author and dates
The rule people break
FAQ schema must match FAQs that are actually visible on the page. Adding questions to your JSON-LD that do not appear in the visible content is a policy violation and can disqualify the whole page from rich results. Same content, both places, always.
4. Core Web Vitals, in practice
Page experience is a confirmed ranking factor, and in India — where most traffic is mobile, often on variable connections — it affects conversions more than it affects rankings.
- Largest Contentful Paint under 2.5s. Usually your hero image. Preload it, set
fetchpriority="high", and make sure it is not a 2 MB PNG. - Interaction to Next Paint under 200ms. Usually caused by heavy JavaScript blocking the main thread.
- Cumulative Layout Shift under 0.1. Set explicit
widthandheighton every image so the browser reserves space before it loads.
The fixes that move the needle most
- Convert images to WebP. A 2 MB PNG photograph becomes roughly 200 KB with no visible difference. On image-heavy sites this alone can halve page weight.
- Lazy-load everything below the fold.
loading="lazy"on every image except the two or three visible on load. - Serve images at display size. A 4000px image rendered in a 400px slot wastes 99% of the bytes downloaded.
- Question background video. A decorative 10 MB loop is the most expensive design decision on most sites.
- Enable gzip or Brotli compression and set far-future cache headers on static assets.
- Audit your JavaScript bundle. Template exports frequently ship a megabyte of interaction code for effects you no longer use.
5. Being citable by AI answer engines
This is the newer discipline — sometimes called Answer Engine Optimisation. The objective is different: not to rank a page, but to be the source a generated answer draws from and names.
Let the AI crawlers in
Check your robots.txt. If it does not mention GPTBot, ClaudeBot, PerplexityBot, Google-Extended and similar agents, their access depends on your default rule — and if you have blocked them, you have removed yourself from AI answers entirely. Decide this deliberately rather than by accident.
Answer the question in the first paragraph
Systems that summarise reward content that states its conclusion early. If a page's key claim only appears in the eighth paragraph after a long preamble, it is far less likely to be extracted. Lead with the answer, then explain it.
Write self-contained factual statements
“We were founded in 2016” is harder to quote than “CNetworQ Private Limited was founded in 2016 in Tamil Nadu, India.” Assume any sentence might be lifted out of context, and make it survive that.
Add an llms.txt
An emerging convention: a plain-markdown file at /llms.txt summarising what your
organisation is, what it does, and the key facts, written for machine consumption. It is cheap to produce
and gives answer engines a clean, unambiguous source instead of making them infer from marketing copy.
Build real FAQ content
Genuine questions with direct answers, marked up with FAQPage schema, are among the most extractable content formats that exist. Write them the way a customer would ask, not the way a brochure would.
The full checklist
| Area | Check |
|---|---|
| Indexability | robots.txt exists and does not block key paths |
| sitemap.xml exists, is current, and is submitted to Search Console | |
| Canonical tag on every page | |
| One hostname; the other 301-redirects to it | |
| On-page | Unique title, 50–60 characters |
| Unique meta description, 150–160 characters | |
| Exactly one H1; logical H2/H3 hierarchy | |
| Descriptive alt text; decorative images use empty alt | |
| Structured data | Organization + LocalBusiness in JSON-LD |
| BreadcrumbList on inner pages | |
| FAQPage matching visible content | |
| Performance | LCP under 2.5s on mobile |
| Images in WebP, sized to display dimensions | |
| Lazy loading below the fold | |
| Compression and cache headers enabled | |
| AEO | AI crawlers explicitly allowed in robots.txt |
| llms.txt published | |
| Answer-first content structure; self-contained factual statements | |
| Measurement | Google Search Console verified and monitored |
| GA4 installed with real conversion events |
Where to start
Verify Search Console first. Not because it is the biggest win, but because without it you are guessing about everything else — and you will have no baseline to prove any of the other work mattered.
Then fix indexability, then titles and descriptions, then performance. Structured data and AEO work last, because they amplify a site that is already sound and cannot rescue one that is not.
