Your report shows that while 99.2% of brands are recognised by name, only 12.4% appear in AI-generated category recommendations. What’s the biggest misconception CMOs have about AI-era brand visibility, and why are they underestimating this shift?
The biggest misconception: CMOs are still measuring “brand strength” using search-era logic in an answer-era world.
For twenty years, brand health lived and died by recall and search share; if 99% of your target market recognized your name and found you on Google when they searched for it, you’d effectively won. Our data shows that instinct is now dangerously misleading. Recognition and recommendation have decoupled. A brand can be a household name- 99.2% of India’s top 100 brands are and still be almost invisible at the exact moment a consumer is asking an AI “what should I buy,” because ChatGPT, Gemini, and Perplexity aren’t retrieving brand names; they’re synthesizing an answer from whichever content they can actually parse, trust, and cite. Only 12.4% of brands are clearing that bar.
Three reasons CMOs are underestimating this:
- They’re grading themselves on the wrong exam. Most CMOs still track share of search and share of voice on social, both of which look healthy. Nobody’s dashboard yet has a “share of AI recommendation” column, so the gap is invisible until you go and test it prompt by prompt, the way our audit does.
- They assume SEO equity transfers automatically. It doesn’t. Being the #1 organic result for a category term doesn’t mean an LLM will cite you in its synthesized answer; AI engines reward structured, extractable, fact-dense content (spec tables, comparison data, clear claims) very differently from how Google’s ranking algorithm rewards backlinks and keyword density. A site built beautifully for search can be almost unreadable to an LLM.
- The failure mode is silent, not visible. When a brand loses search rank, its traffic drops and someone notices. When a brand loses AI recommendation share, nothing breaks; the website still works, the ads still run, the brand simply stops being mentioned in a growing share of purchase-consideration conversations that never touch a search results page at all. There’s no alert for that. It shows up only as a slow erosion in consideration and, eventually, in market share, by which point competitors who invested early have compounded a real lead.
As AI assistants become a discovery channel, what new KPIs should marketers focus on? Do you see “AI share of voice” becoming as important as search rankings?
Yes, and we’d go further: “AI share of voice” won’t just become as important as search rankings, it will eventually eclipse them, because AI answers are increasingly the last stop before a purchase decision, not a stepping stone to ten blue links.
Search rankings tell you whether you’re discoverable. AI share of voice tells you whether you’re recommended. Those used to be roughly the same fight. They aren’t anymore, an AI engine can decide there’s a definitive answer and never send the user anywhere else. If you’re not in that answer, you don’t just lose a click; you lose the conversation entirely.
The KPIs we think marketers need to start tracking, in rough order of maturity:
AI Share of Voice (branded vs. non-branded)
Not “do I get mentioned when someone asks about me by name”; that’s the easy 99% everyone already wins. The number that matters is your share of mentions in category and comparison prompts, where you’re competing against rivals for a recommendation slot you didn’t ask for. This is the single most important new metric, and it’s the one our audits are built around.
Citation Accuracy / Sentiment Alignment
It’s not enough to be mentioned; you need to be described correctly. We regularly find brands cited with outdated pricing, discontinued products, or attributes from a past campaign. An AI engine repeating stale or wrong information at scale is a brand-safety problem hiding inside what looks like a visibility win.
Content Extractability / “GEO readiness”
A more technical, upstream KPI: is your content structured in a way an LLM can actually parse and cite, clear factual claims, comparison tables, spec sheets, FAQ-style structuring, versus prose-heavy brand storytelling that reads beautifully but gives a model nothing concrete to quote. This is closer to a leading indicator than a lagging one; fix it and AI Share of Voice tends to follow.
Prompt-level win/loss tracking
The prompt is becoming the new keyword. Just as SEO teams track rankings for their most valuable search terms, marketing teams will need to track which specific consumer prompts (“best X for Y,” “X vs Z”) they win, lose, or don’t appear in at all and treat closing specific prompt gaps as a working backlog, not an abstract brand health score.
5. Vernacular/regional AI presence
For India specifically, this deserves its own line item. Our data shows branded AI presence in Hindi, Tamil, and Bengali prompts running dramatically lower than English, often near zero, even for brands with strong English-language AI visibility. As voice and vernacular AI usage grows, that gap becomes a real market-share risk, not a nice-to-have.
Based on your research, what are the three most important steps brands should take to become AI-recommended rather than just AI-recognised?
Three steps, in the order we tell clients to actually do them:
Restructure content for machine extraction, not just human persuasion.
This is the foundational fix, and most brands skip straight past it because it’s unglamorous. LLMs don’t reward beautiful brand storytelling; they reward content they can confidently lift a fact from: clear product specifications, honest comparison tables (including where you lose, which paradoxically builds citation trust), FAQ-structured pages that mirror how people actually phrase questions to AI, and unambiguous claims instead of vague positioning language. In our audits, the brands with the highest AI Share of Voice are rarely the ones with the biggest ad budgets; they’re the ones whose websites read like reference documents rather than glossy brochures. If an LLM can’t parse a clean fact out of your page, it will simply cite a competitor’s instead, or a third-party review site, or a Reddit thread, and increasingly, it’s the third option.
Win the third-party layer, because AI engines trust it more than they trust you.
This is the step that surprises CMOs the most. AI engines are deliberately designed to be skeptical of brand-owned content; a company’s own website is treated as a biased source. What gets cited instead is Reddit threads, review aggregators, comparison sites, forums, and independent editorial coverage. That means earned media, authentic user-generated content, and structured third-party reviews are no longer just reputation assets; they’re now the primary raw material AI engines draw on to answer “what should I buy.” Brands that have spent the last decade optimizing owned channels and treating earned/UGC as secondary now need to invert that hierarchy.
Go vernacular and prompt-specific, not just broad and branded.
Being AI-recognised means showing up when someone asks about you directly, in English. Being AI-recommended means showing up in the actual buying-decision prompts, comparison prompts, “best X for Y” prompts, and increasingly, the same prompts asked in Hindi, Tamil, Bengali, and other regional languages, since that’s where a huge and fast-growing share of India’s AI usage already sits, largely unaddressed. Brands need to map the real prompt universe their category gets asked in, not the keyword universe SEO already covers, and build and test content against those specific prompts, the way we do prompt-by-prompt in our audits, rather than assuming general content quality will trickle down into AI recommendation share.
The common thread across all three: this isn’t a marketing campaign you run once; it’s closer to infrastructure you build and then continuously monitor, because AI engines’ training data and retrieval sources shift, and a brand’s position in the answer economy can move underneath them without any warning sign in traditional analytics.
Do you think AI recommendation engines will truly democratise brand discovery, or will larger brands regain their dominance through AI optimisation?
Our honest answer: both forces are real and happening simultaneously. There’s a real democratising window open right now, but it’s structurally temporary, and larger brands are already moving to close it.
The case for democratisation, which is real in the short term:
AI engines don’t currently weight brand size or ad spend the way search and social algorithms do. There’s no equivalent of a paid ranking boost or a follower-count advantage baked into how an LLM decides what to cite. What it rewards instead- clear, structured, extractable content and strong third-party validation- is achievable by a well-run challenger brand with a fraction of a legacy competitor’s budget. Our own data supports this: in several categories we’ve audited, digitally-native and D2C brands are outperforming much larger, better-funded incumbents on AI Share of Voice, purely because their content and reputation infrastructure was built AI-legible from day one, while legacy brands are still running decade-old website architectures optimized for a completely different discovery mechanism. For a genuine window, capital efficiency has been replaced by structural readiness; that’s about as democratic as brand discovery has been in a long time.
The case for re-consolidation, which is where we think this eventually lands:
The advantage smaller brands have right now exists mainly because most large brands haven’t woken up to this shift yet; it’s an execution gap, not a permanent structural one. Once large brands start deploying the same resources they’ve always deployed- bigger content teams, dedicated GEO functions, PR budgets redirected toward the third-party sources AI engines trust, and enough scale to be genuinely present across every regional language and prompt variation- that advantage compounds the same way it always has in owned/earned media. There’s also a training-data effect worth watching: brands with decades of digital footprint, reviews, and citations already have more raw material for models to draw from than a brand that’s five years old, even before anyone does anything deliberate about it.
Where we land: think of this less as “democratisation vs. dominance” and more as a first-mover window that’s open now and closing gradually. The brands winning AI Share of Voice today are disproportionately smaller and faster-moving, but that’s a temporary artifact of large brands’ slow reaction time, not a permanent feature of how these systems work. Our honest advice to challenger brands is to treat this as a genuine, time-limited opportunity to build a moat before the incumbents catch up, and our advice to large brands is that “we’ll always be found because we’re the market leader” is precisely the assumption our report shows is currently failing them.
Over the next three years, what will be the biggest change in market exploration, and how should Indian brands prepare for an AI-first future?
The biggest change over the next three years: the purchase journey will stop starting on a screen full of search results and start starting inside a single AI-generated answer, and for a fast-growing share of categories, it will also end there, with the AI making the actual recommendation and even initiating the transaction.
A few specific shifts we’d flag for Indian brands to prepare for:
The “zero-click, zero-competitor-visibility” purchase will become mainstream.
Today, even when someone searches, they still see ten results and make a choice among them. Increasingly, they’ll ask an assistant once and get one confident answer, often just one or two brand names, not a comparison shelf. The competitive set a consumer even considers is shrinking to whatever the AI decides to surface. Indian brands need to prepare for a world where losing the AI recommendation doesn’t mean losing a ranking position; it means not existing in the customer’s consideration set at all.
Voice and vernacular will drive this faster in India than almost anywhere else.
India already over-indexes on mobile-first and voice-first behavior compared to global markets, and our data shows AI usage growing fastest in Hindi, Tamil, Bengali, and other regional languages, precisely where brand AI-visibility is currently near zero across the board. The India-first AI future won’t look like a Hindi translation of the English AI economy; it’ll be a distinct discovery layer that most brands haven’t even started building for.
Agentic commerce will compress marketing and transaction into one step.
Over the next three years, expect AI assistants to move from “recommend a product” to “add it to cart” or “complete the purchase,” especially as UPI and India’s digital payments rails make agentic checkout technically trivial here. When that happens, being the AI’s chosen recommendation stops being a marketing win and becomes the entire sale; there’s no intermediate browsing stage left for a brand to win back a customer who wasn’t recommended.
Regulatory and platform dynamics will shift who controls the “shelf.”
Just as Amazon and Flipkart became gatekeepers of e-commerce discovery, a handful of AI platforms- OpenAI, Google, and India-specific distribution plays like the Jio–Gemini integration- will become gatekeepers of AI-era discovery. Indian brands should expect this concentration and start building direct relationships and structured-data partnerships with these platforms now, the way smart D2C brands built early Amazon Ads relationships rather than waiting until the channel matured and became expensive and crowded.
How brands should prepare, concretely:
- Treat AI visibility as board-level infrastructure, not a marketing experiment; assign ownership, budget, and a recurring audit cadence, the same seriousness given to SEO a decade ago.
- Rebuild content architecture now, while competitors are still slow, extractable, factual, comparison-honest content compounds, and early movers are already pulling ahead in our data.
- Invest in the third-party and vernacular layers immediately, since both take far longer to build than owned content and both are currently near-empty white space in almost every category we’ve audited.
- Build the internal measurement muscle, track AI Share of Voice, citation accuracy, and prompt-level wins now, before it becomes the default marketing KPI everyone is scrambling to report on.
The brands that treat the next three years as a land grab, the way the smartest brands treated the early days of search and social, will be the ones setting the terms other brands eventually have to compete on.
Enjoyed this interview? Now imagine yours. Write to:
jeevika@thefoundermedia.in
