Kamal Ghorui
Until a few years ago, the internet worked like a city of shopfronts. You could ideally build a website, maybe polish your search rankings, run campaigns, write engaging blogs, and then wait for people to walk in. In essence the discovery was messy, but it was visible. A user searched, scanned ten links on the first page with maybe some sponsored results, compared options, clicked around, and gradually formed a view.
Today, that city is being replaced by something far stranger. As an example, A person no longer types “best MBA college in Delhi”, “which air purifier is worth buying”, or “top logistics company for sustainable supply chains” and patiently sifts through pages of results. Increasingly, they ask a large language model (like ChatGPT, Perplexity etc.) a fuller question in plain English, and expect a clean answer. Just an answer they can trust. Google has expanded AI Overviews and AI Mode in search while OpenAI has pushed deeper into search and product discovery inside ChatGPT. The interface is changing from retrieval to reasoning and from navigation to recommendation.
It is not just an interface change because when discovery moves from search results to synthesized answers, the winner is no longer simply the brand that ranks. The winner is the brand the machine understands, trusts, and chooses to include. Now, that is a very different game. Imagine a parent trying to shortlist schools or a student exploring higher education options or say a family comparing destinations for a holiday. In the old world, they would visit multiple websites reading several pages, open comparison portals, maybe check YouTube or Reddit, then come back later. In the new world, they can ask: “Which institutions in North India are known for industry-linked management education and strong placements?” Or: “Which travel company is best for a luxury cultural holiday in Europe with children?”
The answer they get is not just a list. It is not an answer but rather a compressed judgment.
For twenty years, digital visibility rewarded volume which meant more pages, more keywords, more backlinks, more campaigns. Even when the system was imperfect, the logic was understandable. Marketers could reverse-engineer it. SEO became a discipline and performance marketing became a machine.
Analytics gave the comforting illusion that every click had a traceable origin. LLMs are now disrupting that comfort. When a conversational system answers a user directly, it does not always behave like a search engine. It weighs brand signals, consistency, authority, structured facts, third party mentions, semantic clarity, and contextual relevance. In many cases, the user journey starts further down the funnel because the model has already done the comparison layer on behalf of the user.
Google says AI search is leading people to ask longer and more complex questions, while OpenAI is explicitly positioning ChatGPT as a place where people explore, compare, and decide what to buy.
This means the old question, “How do I rank?” is starting to sound incomplete. The more important question is, “How do I become the answer?” Now, a brand may have a decent website. They may even rank well on Google for traditional keywords. But ask an LLM a relevant high-intent question and the brand may appear weak, inconsistent, confused, or absent. Not because the business lacks value. Because its digital identity was built for indexing, not for inference.
This is the part many executives still underestimate. They think AI discovery is just SEO with a fresh coat of paint but I feel it is closer to reputation engineering for machines. Generative Engine Optimization or GEO in short, is emerging as a serious layer of strategy. GEO is not about stuffing pages with AI buzzwords.
It is about designing a brand so that generative systems can reliably understand who you are, what you do, why you matter, and when you deserve to be cited or recommended. It sits at the intersection of technical structure, semantic clarity, digital authority, content architecture, and brand consistency.
In this case, the economics of discovery are changing too. For much of the web era, marketers optimised for traffic with the visit itself having intrinsic value. More pageviews meant more opportunities to convert, retarget, nurture, and monetise. But answer engines compress that journey. A growing share of user satisfaction may happen before the click. Several observers now describe AI search as a challenge to the pageview economy itself.
It means digital strategy has to evolve from “How do I pull people to my site?” to “How do I shape the machine-mediated perception that happens before my site is even visited?” That is a boardroom question, not just a marketing one. It affects education brands being compared by students. Hospitals being recommended by AI assistants. Hotels being shortlisted by travel planners. Consumer brands being surfaced in conversational shopping. Consulting firms being named in category questions. Even public institutions will increasingly face a new visibility test: are they discoverable in the answer layer of the internet?
The companies that will do well are those that treat AI systems as a new class of audience which will require new instrumentation. Brands now need to know where they appear across LLMs, for which prompts, against which competitors, with what sentiment, under what citation patterns, and in which use-case clusters. This is precisely why a new category of tooling is emerging around LLM visibility intelligence.
The brands still behaving as if a homepage and some search rankings are enough may soon find themselves in the digital equivalent of a crowded marketplace with the shutters down. Still technically there but no longer truly discoverable.
END OF ARTICLE
