SEO, AEO and GEO in India: multilingual search, voice queries and WhatsApp

Short answer: Indian search runs across many languages, happens overwhelmingly on mid-range mobile devices, arrives increasingly by voice, and usually ends in WhatsApp rather than a form. Optimising for a typed English query on a fast desktop connection describes almost none of that, which is why imported playbooks underperform here.

Most search guidance assumes one language, a typed query, a fast connection and a web form at the end. India breaks all four assumptions at once, and a plan built on them will produce rankings that do not convert and content that a large share of the market cannot read.

For the plain definitions of the three disciplines, SEO, AEO and GEO explained covers them. This is about what changes here.

Search happens in more than one language

The IAMAI and Kantar “Internet in India Report 2024”, as reported by PTI, found that 98 per cent of internet users accessed content in Indic languages, and that about 57 per cent of urban internet users prefer consuming content in regional languages. That measures reading and watching, not searching. Our reading, not a measured search figure, is that the gap between what people read and what businesses publish is widest where competition is thinnest.

For English-medium B2B and premium urban services, English still covers most addressable demand. Outside that, an English-only site is invisible to a substantial portion of its market, and the business usually misreads this as low demand rather than as its own absence.

The sensible sequence is not to attempt many languages at once. Prove the model in English, identify the one or two languages that match where your customers actually are, and build genuine content there rather than machine translation. Translated pages that read as machine output are precisely what both search engines and generative systems learn to skip.

The prize is unusual: vernacular content faces far less competition than English content for equivalent commercial intent. A business willing to publish properly in a regional language frequently finds a nearly empty result page where the English equivalent is fiercely contested.

Tamil Nadu shows this most sharply. Tamil-language preference in search there runs stronger than regional-language preference in most Indian metros, which is why our Chennai digital marketing price guide treats Tamil content as a core budget line rather than a later add-on. For how to choose the language and set the pages up, see our guide to regional language SEO in India.

Voice and the shape of the query

Typing in Indian scripts is slower than speaking, which is why voice search is disproportionately common here.

That changes what wins. A voice query is a complete spoken sentence, longer and more conversational than a typed fragment, and it usually expects one direct answer rather than a list of options.

Content built for keyword fragments loses these consistently. Content that states a clear, complete answer in the first paragraph under a heading matching the spoken question wins them. This is the same discipline that earns AI citations, which is why answer-shaped content pays twice in this market.

Justdial, IndiaMART and marketplace listings

Two Indian platforms function as search engines in their own right, and treating them as an afterthought is a common way to lose visibility that never shows up in a Google Search Console report.

Justdial has operated as a local search platform since 1996, reachable by website, app and a national phone number, and its own listing base runs into the tens of millions of businesses across categories. For many local service categories, a Justdial search happens instead of a Google search, particularly among users more comfortable calling or texting a business than filling in a web form. A business absent from Justdial, or present with a stale phone number, is invisible to that entire path regardless of how well its own site ranks.

IndiaMART plays the equivalent role for B2B and manufacturing. It is India’s largest online business-to-business marketplace, connecting a very large base of registered buyers with suppliers across tens of thousands of product categories, and it can generate more enquiries than a supplier’s own website. A supplier profile with complete specifications, certifications and pricing bands functions as a product page that a generative system can cite directly, in much the same way a well-built page on your own site would.

The consistency problem is the one that runs through every part of this work. The business name, address, phone number and category on Justdial and IndiaMART need to match Google Business Profile and the website exactly. A business trading under a slightly different name on IndiaMART than on its own site, or listing an old office address on Justdial after moving, is asking three different systems to agree on who it is when their records disagree. Treat these platforms as citation sources to maintain, not just lead-generation channels to check occasionally.

Tier 1, tier 2 and tier 3 are different businesses

The metros are crowded and expensive. Everywhere else is neither.

Competition in tier 2 and tier 3 cities is a fraction of Mumbai, Delhi or Bengaluru, internet adoption is growing faster, and vernacular search is more prevalent. For many categories the honest analysis is that a business fighting for a metro term would win more, sooner, by targeting six smaller cities properly.

The trade-off is transaction value, which is lower in many categories outside the metros. That makes this a volume play, or a play for businesses that can serve remotely without the economics degrading.

Where local search breaks down inside Indian cities

The tier 1 versus tier 2 framing above is still a national-level view. Inside a single Indian city, local search and AI answers can break down in ways a generic “keep your Google Business Profile tidy” checklist never anticipates, and three specific cases show how differently the problem shows up from city to city.

Noida, Greater Noida and Greater Noida West run three colliding sector-naming systems: Noida numbers its own sectors 1 to 168, Greater Noida abandons numbers for Greek letters such as Alpha and Beta, and Greater Noida West, commonly called Noida Extension, reintroduces numeric sectors that share no relationship to Noida’s own. An address reading simply “Sector 10” is genuinely ambiguous between two different places, and an AI system extracting a location from that text has no way to resolve it unless the page states the distinction itself. Our local AI search guide for Noida covers what that costs a business in practice.

Chandigarh, Mohali and Panchkula behave as one continuous search market to a searcher and to Google Maps, but sit under three separate governments, Chandigarh’s own Union Territory administration, Punjab’s development authority for Mohali and Haryana’s for Panchkula, with no single coordinated plan across them. A listing whose registered state, GST state code and Google Business Profile address disagree with each other, because the business genuinely straddles a jurisdiction most local SEO advice assumes does not exist, loses the trust signals Google’s local ranking depends on. Our guide to local SEO in Chandigarh sets out the fix at the listing level.

Indore splits a different way again: a heritage old-city core around Rajwada and Sarafa Bazaar, a corporate AB Road and Vijay Nagar corridor several kilometres away, and the Pithampur industrial belt selling into an entirely different, B2B audience. A single city-wide profile written vaguely enough to fit all three fits none of them well, which is the specific failure our local SEO guide for Indore walks through.

None of these is the same problem wearing a different city name. One is a naming collision, one is a jurisdictional mismatch, one is a market that was never single to begin with. Generic local SEO advice, built around consistent NAP data and a filled-out profile, does not surface any of the three, because all three sit upstream of the listing itself, in how the city’s own geography and governance are structured. The businesses that get this right treat their specific city’s structural quirk as the first fix, not the profile polish that comes after it. For the listing itself, our guide to local SEO for Indian small businesses covers the address line, verification and reviews.

Mobile constraints are reach constraints

The overwhelming majority of Indian search happens on mobile, frequently on mid-range hardware over variable connections.

This inverts the usual framing of performance. A heavy page does not merely rank lower. It fails to load for a meaningful share of the people who tried, and those people do not report a problem, they leave. Performance here is a reach constraint before it is a ranking factor.

The practical implications are unglamorous and effective: compress aggressively, avoid heavy frameworks where static output would do, keep the critical content in the initial payload, and test on a mid-range device on a throttled connection rather than on office broadband.

WhatsApp sits between the click and the enquiry

This is the piece imported playbooks miss entirely.

A large share of Indian buyers who find you through search want to continue on WhatsApp rather than fill in a form or make a call. A site that offers no such route loses them at the exact moment they were ready to talk.

It has no direct effect on ranking. It has a decisive effect on whether ranking becomes revenue, which is the only reason ranking mattered. A visible WhatsApp path from the pages that attract search traffic, with a real person or a properly built automation behind it, converts traffic that a contact form would have lost.

Consent still applies. Under the DPDP Act, the messages that follow are governed by how you obtained permission, so the opt-in has to be genuine.

Bulk SMS needs a registration step WhatsApp does not

The WhatsApp route above assumes consent-based messaging through an opt-in. Promotional SMS in India runs under a separate and stricter regime that catches businesses who assume it works the same way as email.

Under the Telecom Commercial Communications Customer Preference Regulations, 2018, the Telecom Regulatory Authority of India requires any business sending bulk commercial SMS to complete a three-layer registration on a blockchain-based Distributed Ledger Technology platform: the business entity itself, each sender ID it messages from, and every distinct message template it plans to send, all registered and approved before a single message goes out. TRAI’s August 2024 directions go further, requiring operators to block messages carrying links or callback numbers the sender has not whitelisted and to reject messages whose delivery chain cannot be traced, which means an unregistered campaign does not arrive late, it does not arrive at all.

This matters for a business that has invested in ranking or earning an AI citation and then tries to close the resulting leads with an SMS drip sequence. A plan that budgets for content and search but treats SMS as a same-day add-on discovers the registration step at the worst possible moment: mid-campaign, with leads going cold while templates sit in an approval queue. WhatsApp Business messaging runs on a different, opt-in-based model and sits outside DLT entirely, which is one more reason the two channels are not interchangeable even though both look like bulk messaging from the outside.

Hreflang and schema for Hindi and regional pages

Publishing in Hindi or a regional language solves half the problem. The other half is telling search engines which page belongs to which language and region, and that is what hreflang and structured data do.

Google’s documentation for localized pages builds hreflang values from an ISO 639-1 language code, optionally combined with an ISO 3166-1 region code, so a Hindi page for Indian users takes the value hi-IN, a Tamil page takes ta-IN, and an English page written for the Indian market specifically, as distinct from a UK or US version, takes en-IN. Each language version needs a complete set of hreflang link elements (rel equals alternate) in its page head, one for every language variant including itself, and the sets need to match exactly across all versions. A Hindi page that points to the English version, where the English version does not point back, is a broken reference that Google’s own documentation says will be treated as unreliable. Google also recommends an x-default version for visitors whose language or region does not match any of the alternates.

The separate-URL-versus-section decision matters more in India than in a single-language market. A genuinely separate URL structure, such as a language subdirectory, lets each version carry its own title tags, meta description and internal links written for that language’s actual phrasing, not a template with swapped words. A translated section bolted onto an English-first structure tends to inherit English-shaped headings and URL slugs that read as foreign to a Hindi or Tamil searcher, undermining exactly the credibility a regional-language page is meant to build.

Schema markup on these pages should declare the content’s language explicitly through the inLanguage property alongside whatever schema type the page already uses, such as LocalBusiness or Service. That gives a generative engine answering a Hindi-language question a structured signal, not just prose, confirming this specific page is the Hindi answer rather than a machine-translated copy of the English one.

The DPDP Act: capture, not content

The Digital Personal Data Protection Act 2023 sets consent and notice requirements for collecting and processing personal information.

Its effect on this work is entirely on the capture layer: forms, WhatsApp opt-ins, stored contact data and how you describe your handling of it. It places no restriction on what you publish, so it does not constrain SEO, AEO or GEO content in any way.

Obligations depend on your specific circumstances, so take advice rather than relying on a marketing guide for a legal question.

What Indian AI answers draw on

Answers to Indian commercial questions lean on government and regulator sources, established Indian publications, large marketplaces and directories with domestic volume, and businesses whose own pages state facts plainly enough to quote.

Indian-context sources dominate for Indian queries. A modest set of credible domestic mentions does more than a much larger international profile, which is the reverse of what businesses with global ambitions tend to assume.

Your own pages carry the other half. State your service area, your pricing structure, your qualifications and your process as plain facts. Promotional prose does not get quoted because there is nothing quotable in it. Our AEO guide for small businesses in India turns this into a checklist.

Hiring for AEO in India: an agency, a tool, or neither

Our view: most Indian small businesses should run the free checks first and pay an agency only for work they cannot do themselves. A relabelled SEO retainer is not AEO.

Start with what costs nothing. Google says “There are no additional requirements to appear in AI Overviews or AI Mode”, and that no other special optimisation is necessary (source). Traffic from those AI features is counted inside the Search Console Performance report under the Web search type, so it is not reported separately there.

Bing is more useful here. In February 2026 Bing Webmaster Tools added an AI Performance report in public preview, showing how often your pages are cited in Microsoft Copilot and Bing’s AI summaries. Bing says the figures are a sample and do not indicate ranking, so read them as a signal, not a score.

Then test by hand. Ask the same buyer questions in ChatGPT, Gemini and Perplexity, in English and in your customers’ language, and record who is named. That is our method, and it costs an afternoon.

What a paid AEO tool adds. In our view, repeated prompt tracking over time, so you can see whether a mention appears or disappears. It does not fix anything on its own. If you have nobody to act on what it shows, it is a dashboard, not a solution.

When an agency is worth it. When the gap you found is regional-language content, structured data, or listings consistency across Google Business Profile, Justdial and IndiaMART, and you lack the time to do it. Quotes for marketing services in India are usually written before 18 percent GST, so ask whether the figure includes it. For price ranges, see what SEO costs in India.

What to ask before signing.

• Which prompts will you test, and in which languages?
• Will you share every change you make to my site? Google’s own hiring guidance suggests asking this.
• What results do you expect, in what timeframe, and how will you measure them?

Google says “No one can guarantee a #1 ranking on Google” (source). The same logic applies to anyone promising a ChatGPT or AI Overview mention.

If you want a second opinion on where the gap is, see our SEO, AEO and GEO service.

Sequencing on an Indian budget

Fix mobile performance first. It is a reach problem, it affects everything downstream, and it is usually the cheapest large gain available.

Add the WhatsApp path early. Conversion work before traffic work, because traffic into a funnel that loses ready buyers is wasted spend.

Publish answers to real questions in English. Fast to produce, quick to earn citations, and it validates which topics actually generate enquiries.

Then extend into one regional language. Choose by where your customers are, not by speaker numbers nationally. Write, do not translate.

Then build city depth in tier 2 and tier 3 markets where the competition is thin, but only with genuine local substance.

If you are paying someone to do this, our guide to what SEO costs in India explains how to price each step as work delivered rather than keywords promised.

Measuring when volume is large and intent is thin

Indian search volumes are high enough to flatter a report while producing very little business.

So measure enquiries rather than traffic, and segment by language and by city tier rather than nationally. A blended national figure will hide both the vernacular market you are absent from and the metro traffic that never converts.

Watch the WhatsApp path specifically. If search traffic rises and conversations do not, the problem is the handoff rather than the ranking.

Then test AI answers directly in both English and your target language, because the results differ sharply and the competitor named in one is frequently not the one named in the other.

Where Indian businesses go wrong

• English-only by default. Then reading their own absence from vernacular search as an absence of demand.
• Machine-translated vernacular pages. Worse than none, because they consume budget and get skipped by both engines and readers.
• Desktop-built sites. Tested on office broadband, unusable on the devices most of the market actually holds.
• No WhatsApp route. Ranking achieved, buyer lost at the last step.
• Fighting for metro terms while six tier 2 cities sit nearly uncontested.

Frequently asked questions

Should I hire an AEO agency in India or use an AEO tool? Run the free checks first: Search Console, Bing’s AI Performance report, and manual prompt tests in English and your customers’ language. A tool tracks prompts over time but fixes nothing by itself. Pay an agency for the regional-language content, structured data or listings work you cannot do yourself, and be wary of anyone guaranteeing an AI mention.

Do I need content in Hindi and regional languages to compete in India? It depends on who buys from you. For English-medium B2B and premium urban services, English alone still covers most of the addressable demand. For consumer services outside the largest metros, English-only content leaves a large and growing share of searchers unreachable, because nearly all Indian internet users already read and watch content in Indian languages. The efficient approach is to prove the model in English, then extend into the one or two languages that match where your customers actually are, rather than attempting many at once.

How does voice search change what I should publish? It changes the shape of the query. Voice searches are longer, conversational, and phrased as complete questions rather than keyword fragments, and they are disproportionately common in India because typing in Indian scripts is slower than speaking. Content that answers a full spoken question directly, in one clear paragraph near the top, wins these. Keyword-fragment pages built for typed search consistently lose them.

Why does WhatsApp matter for SEO in India? Because it usually sits between the click and the enquiry. A large share of Indian buyers who find you through search will want to continue the conversation on WhatsApp rather than a form or a phone call, and a site that offers no such path loses them at exactly the moment they were ready. It does not affect ranking directly. It affects whether ranking turns into business, which is the only reason the ranking mattered.

Does the DPDP Act change how I do marketing? It changes how you handle personal data, not what you publish. The Digital Personal Data Protection Act 2023 sets consent and notice requirements around collecting and processing personal information, which affects your forms, your WhatsApp opt-ins and your stored contact data. It places no restriction on SEO, AEO or GEO content. Take specific advice on your obligations rather than relying on a marketing guide.

Are tier 2 and tier 3 cities worth targeting? Often they are the better opportunity. Competition is a fraction of what it is in Mumbai, Delhi or Bengaluru, internet growth is faster, and vernacular search is more prevalent, which means a business willing to publish in the local language faces very little competition. The trade-off is lower average transaction value in many categories, so it works best where volume compensates or where you serve remotely.

How much does mobile performance actually matter here? More than in most markets. The overwhelming majority of Indian search happens on mobile, frequently on mid-range devices and variable connections. A heavy page does not just rank worse, it fails to load at all for a meaningful share of your audience. Performance in India is a reach constraint before it is a ranking factor.

What do Indian AI answers cite? Government and regulator sources, established Indian publications, large marketplaces and directories with Indian volume, and businesses whose own pages state facts plainly. Indian-context sources dominate for Indian queries, so a modest set of credible domestic mentions outperforms a larger international profile.

How long does this take in India? We will not promise a timeframe. Google’s own advice is to ask anyone you hire what results they expect, in what timeframe, and how they measure it. In our view, changes on your own pages can show up in AI answers sooner than rankings move, but nobody controls when. Vernacular pages often face a thinner field, which is our reading rather than a measured figure.

Do I need a separate URL for each Indian language, or can I use one page with a language switcher? A separate URL for each language, correctly marked up with hreflang tags such as hi-IN or ta-IN pointing to each other, works far better than a single page that swaps text with a switcher. Search engines index URLs, not toggled states, so a language switcher alone leaves the non-default language largely invisible to search.

Why do AI answers sometimes place my business in the wrong part of an Indian city? Usually because the city’s own geography or governance is more fragmented than an AI system’s generic address parsing accounts for. Noida, Greater Noida and Greater Noida West run three colliding sector-naming systems; Chandigarh, Mohali and Panchkula are one search market split across three governments; Indore splits into a heritage core, a corporate corridor and an industrial belt with almost nothing in common. A page that states its exact location, jurisdiction and service area in plain language resolves the ambiguity an AI system otherwise has to guess at.

Do I need DLT registration if I only message customers on WhatsApp? No. DLT registration under the Telecom Commercial Communications Customer Preference Regulations applies specifically to bulk commercial SMS sent through telecom operators, and TRAI’s August 2024 directions also require operators to block messages carrying links or callback numbers the sender has not whitelisted. WhatsApp Business messaging runs on a separate, opt-in-based model outside that regime. The two are easy to conflate because both are bulk messaging channels, but only one of them needs registration before you can send a single message.

---

Want to know what AI engines say about your business in English and in your customers’ language? Get a free SEO and AI visibility audit.