Your brand has an AI reputation, you’re probably not tracking it: Pulp Strategy’s Ambika Sharma

NEW DELHI: There is a sentence being written about your brand right now. A machine is composing it, your customer is reading it, and the odds are overwhelming that you have never seen it. That is the problem Ambika Sharma decided to solve.

Sharma is the founder and chief strategist of Pulp Strategy Communications, a New Delhi-based firm with 24 years of brand work and 124 international awards behind it. She is also, as of 2025, the architect of NeuroRank, a patent-pending AI visibility intelligence platform that monitors what ChatGPT, Gemini, Claude, and Perplexity are actually saying about brands. Not what those brands wish the machines would say. What they are, in cold fact, saying today.

To understand why she built it, you have to understand what she saw happening to the question every marketer used to ask. For two decades, marketing ran on a single anxiety: where do I rank on Google? Sharma thinks that question is heading for the archive.

“The buyer no longer scrolls a page of links,” she says. “They ask a machine and take its answer.” The numbers make her point for her. About 60 per cent of searches now end without a click. The website, once the coveted destination at the top of the funnel, has been quietly demoted to a source the AI may or may not choose to cite.

This is the terrain that Sharma calls LLMO, or Large Language Model Optimisation, also known as GEO (Generative Engine Optimisation). It is not a tweak on top of SEO. It is a different discipline altogether, concerned not with where you rank on a results page, but with whether a model recommends you, ignores you, or, most damaging of all, describes you incorrectly to a customer who has no reason to question the answer.

“In five years, this stops being a specialism and becomes the centre of marketing,” she says. “Because the answer is where the decision now happens.”

That shift, however, is arriving in a landscape where most brands are not merely unprepared but actively misinformed about where they stand. Ask Sharma about the biggest mistakes brands are making and she will give you two, both of them quietly devastating.

The first is misplaced confidence. Most marketing teams have typed their brand name into ChatGPT and been broadly satisfied with what came back. Sharma has a name for the problem with that approach: logged-in bias. The platforms know you. They have been trained on your emails, your presentations, your documents. They will be pleasant to your face. They are not being nearly as pleasant to your customers.

“It often omits the brand, hands its strength to a competitor, or states something outdated as fact,” she says. “Most brands have never read the sentence or the sentiment the machine writes about them.”

The second mistake is paralysis. Only around 14 per cent of marketers track whether AI engines actually cite them. The rest are, in her words, flying blind into the channel that increasingly decides the purchase.

What makes the stakes especially high is the architecture of AI answers. There is no page two. Citations concentrate in a short list. If a brand is not in that list, it is not ranked low. It is simply absent from the moment the decision is made. “When we run a NeuroRank diagnostic,” Sharma says, “the first shock for most CMOs is simply seeing what the model says. They have never looked. It is not possible to see manually by running 10 or 20 queries.”

That gap between assumption and reality only deepened as Sharma built and stress-tested the platform. Across more than 700 brands studied, NeuroRank consistently found that ChatGPT, Gemini, Claude, and Perplexity do not agree with each other. The same brand gets described differently by each model, because each is trained on different sources and weights them differently. One model may cite your owned content. Another leans heavily on third-party reviews. A third surfaces a competitor in the same breath as your name.

“A brand can be visible in one engine and invisible in another, and most never realise the gap exists,” she says.

This discovery dismantled the SEO playbook entirely. In search, you optimise for one algorithm. In the AI era, you are conditioning multiple models, each with its own logic, drawing from sources you do not own. Sharma spent 150 CMO meetings presenting NeuroRank results before the product launched in 2025, precisely because the variation between engines is the real problem. “Monitoring one is hard,” she says. “Monitoring all of them at scale is harder. Governing all of them is the really hard part that moves the result.”

If the problem is structural, the fix has to be too. Which means, for most brands, a fairly uncomfortable rethink of what their content is actually for. Sharma is direct about what needs to change, and it goes well beyond adding keywords or tweaking meta descriptions.

“Stop writing just for a human to skim or an algorithm to rank,” she says, “and start writing with correct structure for a model to trust and cite.” The model rewards clear, structured, factual, well-sourced content it can lift and attribute. It punishes vague brand fluff. Your content’s job is no longer to hold a reader on the page.

Consumer behaviour has moved on and Sharma thinks brands should move with it rather than against it. “Your role as a brand is now that your answers, the ones AI models are dishing out with or without your effort, must be the most accurate, most citable account of your category.” The brand that treats its website as a source of truth for machines rather than a brochure for humans, she argues, wins the citation, the brand inclusion, and ultimately the customer.

The commercial argument is straightforward. AI-referred traffic converts at substantially higher rates than traditional organic, because the buyer arrives already informed and closer to a decision. Fewer visitors, better ones.

Sharma has spent more than two decades in brand strategy and has navigated the upheavals of search and social from the inside. She is precise about how this shift differs from what came before, and the distinction matters.

“Traditional digital marketing was about controlling your own channels. Your site, your ads, your pages, your influencers. You owned the surface and you optimised it. AI marketing is about influencing a surface you do not own.”

The model builds its answer largely from sources outside your control, then speaks to your customer on your behalf. You cannot buy the top slot in an AI answer, and when paid positions do eventually arrive, customers are likely to treat them with the same suspicion they reserve for paid search results. You earn the citation by being the brand the model has the most reason to trust.

“The discipline that took us through the rise of search and social prepared us for one thing above all: the steep learning curve every platform shift creates,” she says. “AI visibility is the next curve, and it is steeper and longer than the ones before it. The brands that learned search early won search. The same window is open now, and it will not stay open.”

For anyone still inclined to treat this as tomorrow’s problem, Sharma offers some numbers that have a habit of ending that conversation. When asked whether AI visibility will eventually matter as much as SEO and social media management, she suggests it already does.

“It is simple maths. Twenty-five per cent of all search has already moved to AI. That includes 51 per cent of all B2B software buyers and 78 per cent of investors globally. Premium users migrated first. This 25 per cent is already your economy-churning audience.”

The traffic growth figures she cites are, frankly, the kind that stop rooms. Retail up 693 per cent. Travel up 539 per cent. Financial services up 266 per cent. Technology and software up 120 per cent. These are the numbers for AI-driven traffic growth across industries during the 2025 festival season. The budgets are following: around 94 per cent of CMOs plan to increase GEO investment this year, and the category is projected to grow from under a billion dollars to tens of billions within the decade.

There is also a hidden tax for those who wait. Paid search budgets are increasingly being eaten by a phenomenon called query fanout, where the AI expands a single search into multiple related queries. Sharma says this is burning 30 per cent or more of budgets on ads that are not converting, because the brand has not established trust with the underlying models.

“The brands treating AI visibility as optional are making the exact mistake brands made about SEO in 2005 and social in 2010,” she says. “Only this one will hit harder or reward better.”

One of the more counterintuitive threads running through her argument is that smaller brands and startups stand to gain more from AI visibility optimisation than large enterprises. 

She says “On Google, a startup competes against enterprises with twenty years of domain authority and huge link budgets. It almost always loses on size. The AI answer does not work that way.” The model cites the source with the clearest, most accurate, most relevant information. Budget does not buy clarity.

The data from G2 reinforces the point. A third of B2B buyers have now purchased from a vendor they had never heard of before an AI surfaced it. The machine, in other words, is already making introductions that the old marketing hierarchy would never have permitted. “For a smaller brand willing to be genuinely useful,” she says, “that is the best opening in a generation.”

The temptation, when confronted with a new platform challenge, is to run a one-time audit, fix the obvious problems, and consider the matter closed. Sharma is clear that AI visibility does not work that way, and the reasons are structural rather than arbitrary.

“It is a position you hold.” Models retrain. Sources shift. Competitors begin conditioning their own presence. A brand cited well this quarter can quietly disappear from the next quarter’s answers and never know, because no dashboard it owns tracks the response. Absence also compounds. Models learn from what they already say, so early invisibility becomes entrenched invisibility.

This is why NeuroRank is built around what Sharma calls Model Preference Engineering, a continuous governance model rather than a one-time fix. The results from a 90-day financial services engagement are instructive: sustained conditioning lifted a brand’s AI visibility by 30 per cent and its citation frequency by 12 per cent across all four major engines.

“That does not come from checking once,” she says. “It comes from holding the position.”

Pull all of it together and a picture emerges of what separates the brands that will thrive in the AI era from those that quietly fade from the conversation. Sharma distils it to three principles, and notably, none of them is budget.

The first is seeing early and acting quickly. The brands that win will be the ones who looked at what the AI says about them before their competitors thought to look. Most have not looked properly. That window of competitive advantage is still open.

The second is substance over spin. You cannot flatter your way into a citation. You earn it by being the clearest, most accurate, most useful source in your category. Years of random content accumulation have cluttered the house, and that clutter harms more than most brands realise. Trial and error in AI conditioning is not a neutral exercise.

The third is treating this as a discipline rather than a campaign. The winners will govern their AI presence continuously, the way they once managed SEO and social, rather than running one audit and hoping the world stays still.

“When AI tells your customer your story, is it telling the truth? Is it even telling the customer anything about you? The brands that can answer that question confidently will own the next decade. The ones that cannot will spend it wondering where their customers went.”