Proprietary research · 50 reports
India's MSME AI-Visibility Gap
Buyers in India increasingly ask an engine and act on the answer. This strand reads one question across the Indian small-business economy: when someone asks ChatGPT, Gemini, Google's AI Overviews, or Perplexity a buying question, do real MSMEs get named, or does the answer default to an aggregator, generic advice, or English when the buyer typed in Hindi. It is built from public data across eight themes, from ONDC's open protocol to the vernacular answer gap to the zero-click economy, and it holds to one rule: where a measurement does not yet exist, it says so and sets out how one could be run.
5 reports
The Indian Share-of-Answer Baseline
Whether India's MSMEs appear inside AI answers at all, a measurement framework, and what the published evidence already implies.
The Share-of-Answer Index: Are India's MSMEs Even in the Room When AI Engines Answer?
No public dataset measures whether Indian MSMEs appear inside AI answers. Here is a framework to measure it, and what the evidence already implies.
Read the deep-dive100 Million Weekly Users, How Many Local Answers? Testing ChatGPT's India Scale Against Its India Depth
India is ChatGPT's second-largest market by user count. That counts people opening an app, not whether it names a real local business when one is asked for.
ReadGEO Isn't SEO for Bharat: Do Aggregators Eat the Answer Before the MSME Gets a Turn?
Applying the GEO framework and citation data to ask whether IndiaMART, JustDial, and Sulekha out-cite individual MSMEs in AI answers, and how to test it.
ReadThe Category League Table: Which MSME Verticals Actually Get Named by AI, and Which Don't
No public ranking shows which Indian MSME categories AI names a real business in, versus generic advice. The method, and what the evidence implies.
ReadThe 1.6% Problem: What It Means That Almost No MSME Has a Working Website
One audit finds 1 percent of MSMEs have a working website, a survey finds 19 percent self-report one. Either way, formal India is mostly unreadable to machines.
Read3 reports
Who the Engine Names: Sovereignty, Geography, and Exports
A sovereign India-built answer layer, the metro bias, and who gets named when a foreign buyer asks an AI for an Indian manufacturer.
Will India Get Its Own Answer Engine? Bhashini, Sarvam, and the Sovereign AI-Visibility Stack
Bhashini has crossed six billion language requests and Sarvam is building sovereign Indian models. What would an India-built answer layer mean for MSME visibility?
Read the deep-diveThe Export Answer Gap: When a Foreign Buyer Asks AI for an Indian Manufacturer, Who Gets Named?
MSMEs generate nearly half of India's exports. When AI answers a buyer's sourcing question, does it name the exporter, or the marketplace that lists them?
ReadMetro Privilege: Does an MSME Need to Be in Mumbai or Delhi to Be an AI Answer?
A proposed method for testing whether AI answers name real Tier-2 and Tier-3 businesses or default to generic advice, read against the published evidence.
Read4 reports
The Open Catalog Advantage
Why an ONDC listing on an open protocol may outlast a closed-platform listing once AI agents do the finding and choosing.
The Open Catalog Advantage: Why an MSME's ONDC Listing May Outlast Its Amazon Listing as AI Agents Start Shopping
An Amazon listing and an ONDC listing look alike to a shopper, but only one is built on an open protocol a shopping agent can read without permission.
Read the deep-diveThe Rs 277 Crore Bet: What the MSME-TEAM Scheme's Catalogue Money Is Actually Buying, and Whether It's AI-Ready
The government pays by the SKU to build ONDC catalogues for half a million MSMEs. Whether that data is ready for AI shopping agents is a separate question.
ReadThe Multicast Difference: Why an ONDC Catalog Entry Isn't Competing for a Search Slot the Way an Amazon Listing Is
On Amazon, a seller competes for a slot in one ranked index. On ONDC, the protocol multicasts every matching search to the seller. Not the same contest.
ReadReading ONDC's Public Disclosure Filings: What 70+ Seller and Buyer Apps Reveal About How They Actually Rank MSME Listings
ONDC requires every buyer and seller app to publish how it ranks listings. Most sellers have never read a filing, and some filings are barely readable at all.
Read5 reports
Agentic Commerce Protocols and the ONDC-Only MSME
AP2, ACP, and Beckn: the three protocols that will decide whether an Indian MSME can be bought by an AI agent.
AP2, ACP, Beckn: A Reader's Map of the Three Protocols Deciding Whether an Indian MSME Gets Bought by an AI Agent
Google's AP2, the OpenAI and Stripe Agentic Commerce Protocol, and ONDC's Beckn layer each solve a different piece of agent buying, and today they do not talk to each other.
Read the deep-diveBeckn as Contract Language for Machines: What It Means for an MSME When the "Buyer" Is an AI Agent, Not a Person
Beckn's architects describe it as machine-readable contracting infrastructure for a future where an AI agent, not a person, does the browsing and buying.
Read45 Minutes vs 4 Months: What Beckn Onix's Onboarding-Speed Claim Means for How Fast New AI-Native Networks Could Route Around ONDC
Google Cloud says a Beckn network that took months to build now takes about 45 minutes, a real question for MSMEs that treat ONDC as their only channel.
ReadDoot Meets ONDC: Does India's Proposed Citizen AI Agent Actually Have a Commerce Layer to Shop On?
A text search of the Digi Doot whitepaper finds zero mentions of ONDC. Its GitHub spec names ONDC twice, then omits it from its own comparison table.
ReadTwo Roads to the Same Shopper: What Happens When Perplexity's Agentic Checkout Meets an ONDC-Only MSME
AP2, ACP, and UCP wire AI shopping agents into closed checkouts. No official bridge yet connects them to ONDC, the open network carrying India's MSMEs.
Read8 reports
The Vernacular Answer Gap
Nine hundred million Indians search in Indic languages while the engines answer in English. Script, code-mixing, and the language stack.
The Vernacular Answer Gap: Why 900 Million Indians Search in Indic Languages While AI Engines Answer in English
India searches in 22 languages, but AI engines answer best in English. Why Indic-language answers degrade, and what the gap means for small businesses.
Read the deep-dive22 Scheduled Languages, One Answer Engine: Mapping Where Google's AI Overviews Actually Work in India
Google's AI answers launched in India with Hindi only. Current Google pages show 13 of 22 scheduled languages covered; 9, roughly 51 million speakers, are not.
ReadBhashini's Billions of Calls, Zero MSME Directory: What India's Language Stack Wasn't Built For
Bhashini moves billions of translation calls across 22 languages, but it was built for citizen services, not to make MSMEs legible to AI answer engines.
ReadThe Script Problem: Devanagari, Roman Hindi, and the MSME That Doesn't Exist Twice
A business written in Devanagari and the same business written in Romanized Hindi look, to most AI crawlers, like two different entities rather than one.
ReadThe Hinglish Blind Spot: How Code-Mixed Queries Confuse AI Answer Engines
Romanized Hindi-English, the default typing mode for most Indian mobile users, falls into a text-normalization gap the models behind AI answers still misread.
ReadThe WhatsApp Business Language Ceiling: 11 Languages, 22 Scheduled, and the MSMEs Left Off Both
Cross-referencing Meta's own supported-language list against India's Eighth Schedule shows WhatsApp Business natively covers ten of twenty-two scheduled languages. Twelve get nothing.
ReadWhich Indian Language Does ChatGPT Actually Trust? A Reading of the Hallucination Data
A 2025 benchmark of six LLMs found Hindi replies scored far worse on semantic accuracy than Mandarin, a gap every vernacular-reliant MSME should know.
ReadWomen-Led MSMEs and the Vernacular Interface Gap
Udyam counts 3-plus crore women-led enterprises. A 2025 survey found their strong vernacular-interface preference, naming who the gap exposes.
Read9 reports
The Udyam-to-Visible Gap
Formalizing tens of millions of MSMEs into Udyam did not make them findable. Registration is not machine-legibility.
The Udyam-to-Visible Gap: Inside the 7.8 Crore Registrations AI Engines Cannot See
India formalized 78 million MSMEs into Udyam in under six years. No public dataset yet measures how many are findable when an AI engine is asked a buying question.
Read the deep-diveThe Credit Visibility Correlation: Are AI-Discoverable MSMEs More Likely to Get a Loan
Lenders already score digital footprints. No dataset yet tests whether AI visibility moves loan odds. Here is what the published evidence implies.
ReadFormalize First, Get Found Never: Which Udyam Sectors Are Most Invisible to AI Search
Trading leads Udyam registration at 42.89%. No dataset yet ranks sectors by AI-answer visibility. Here is what the evidence implies, and how to measure it.
ReadGST Says 1.5 Crore, Udyam Says 7.8 Crore: What the Six Crore Gap Says About Findability
About 1.5 crore Indian businesses hold active GST registrations against 7.83 crore Udyam registrations. The gap is real, but not the headcount it looks like.
ReadThe Instagram Illusion: Formally Registered but Structurally Invisible to Generative Engines
Millions of Indian MSMEs treat Instagram or WhatsApp as their digital presence, a channel AI answer engines are built, by their own rules, not to read.
Read7.83 Crore Registrations, Zero Visibility Metric: Why No One Tracks Whether MSMEs Can Be Found
India counts Udyam-registered enterprises down to the decimal on credit, jobs, GDP share, and exports. No government report measures whether they can be found.
ReadReserved for MSEs, Invisible to AI: What the 25% Procurement Mandate Reveals About Discovery
India reserves 25% of government procurement for small enterprises by law. The gap between the mandate and who wins orders previews the AI-visibility problem.
ReadThirty Percent of GDP, Absent From the Answer Layer: The Economic Cost of MSME AI Invisibility
MSMEs generate about 30 percent of India's GDP and nearly half its exports, yet little verified data shows how visible that weight is to the AI answer layer.
ReadThe Last Mile Formalization Trap: A Udyam Number Is Not a ChatGPT Answer
Government schemes treat a Udyam certificate as the finish line of going digital. Being the answer an AI system gives a buyer is a separate race, and nothing in the formalization pipeline runs it.
Read8 reports
India's Discovery Stack and Its Machine-Blind Channels
WhatsApp, Justdial, Maps, and a near-total Google monopoly: where Indian buyers discover, and where AI answers now intercept.
The WhatsApp Blind Spot: India's Biggest Discovery Channel Is Invisible to AI Answer Engines
WhatsApp is where most online Indians already chat with businesses, yet its catalogs and conversations sit behind a login wall no AI crawler can read.
Read the deep-dive56 Million Justdial Listings, 11.5 Million Udyam MSMEs: Mapping India's Discoverability Gap
Justdial counts 56 million listings. Udyam counted about 11 million MSMEs in 2022, over 78 million now. Neither number says who AI engines can find.
Read91% of India Chats With a Business Every Week. Zero Percent of That Conversation Trains an AI.
WhatsApp carries more real buyer questions than any channel in India, answered correctly every day, and almost none of it ever reaches an AI engine.
ReadJustdial's Missing Machine Layer: 160 Million Reviews No AI Engine Can Read
Justdial holds 160 million phone-verified reviews, one of India's largest trust ledgers. Its own robots.txt blocks OpenAI's crawlers from most of it.
ReadThe Map Pack Paradox: Winning Google Maps in India No Longer Means Winning the Answer
AI Overviews now sit above the local three-pack for a growing share of India queries. A top-ranked local listing can still lose the answer a buyer reads.
ReadRural India's Internet Growth Is Outrunning Its AI-Readable Footprint
Rural India added internet users nearly four times faster than urban India, but its MSMEs still sell mainly by phone and WhatsApp, not machine-readable channels
ReadThe 78% Statistic Nobody Can Source: What We Actually Know About Indian SMBs and WhatsApp
A number everyone repeats and no one can source. What the traceable Meta, Kantar, and government data on Indian SMB WhatsApp use actually says, and does not.
ReadThe 97% Country: What Google's Near-Total Search Monopoly in India Means Once Google Also Owns the Answer
Google holds roughly 97% of India's search market and now writes the AI summary above it, a gatekeeper risk sharper for MSMEs than in most large economies.
Read8 reports
The Zero-Click Answer Economy in India
India ranks among the world's highest for AI Overviews prevalence. What happens to a small business when the answer box keeps the click.
The Zero-Click Economy Arrives in India: Inside the 83% of AI-Answered Searches That Never Reach an MSME's Page
Queries that trigger an AI Overview end without a click 83% of the time. Against India's AI-adoption curve, that reshapes how a small business gets found.
Read the deep-diveAI Mode Goes Hindi: What a Second Zero-Click Language Layer Means for MSMEs That Never Answered in English
AI Mode now answers in Hindi and 7 more Indian languages, a surface most MSME content, built almost entirely in English, was never positioned to be cited in.
ReadThe Answer Box Ate the Click: Mapping Google's AI Overviews Rollout in India Since August 2024
From a six-country launch in August 2024 to Hindi AI Mode in 2025, a sourced timeline of how fast the zero-click answer layer grew under Indian MSMEs.
ReadFourth in the World: What India's 26.8% AI Overviews Prevalence Rate Means for a Small Business's Web Traffic
India ranks 4th of 50 countries for AI Overviews prevalence, at 26.8%, ahead of the US, UK, and Canada. What that share means for an MSME's web traffic.
ReadThe Jio-Gemini Default: When 500 Million New Users Never Learn to Click at All
Google made Gemini the free default for Jio's 500 million subscribers. Whether that cohort forms a click-through search habit is an open, unmeasured question.
ReadThe Missed-Call Economy Was Zero-Click Before AI Was: What Justdial's Two Decades Predict About the AI-Answer Shift
Justdial built a phone-first, ask-and-get-an-answer habit in India decades before generative AI. The habit is documented; who it favors next is not.
ReadTraffic Without a Trace: Why an MSME's Analytics Dashboard Can't See the Zero-Click Loss
Standard analytics count sessions. A zero-click AI answer never creates one, so a flat traffic chart can hide demand a business never knew it lost.
ReadZero-Click by Design: Why WhatsApp's 535 Million Indian Users Never Generated a Single Indexable Click
India's largest commerce channel was built to keep the entire transaction inside a chat window years before AI Overviews made zero-click a search-industry worry.
ReadWhere does your business actually stand?
This strand measures the market. A Machine-Readiness Score measures you: a specialist-reviewed read of where a business stands across classic search, the local map pack, AI answers, and reputation. No guaranteed number, and no obligation.