Lauren Wiener, Managing Director and Senior Partner at BCG, and Parul Bajaj, Managing Director & Partner and India leader - Marketing, Sales & Pricing at BCG India, discuss how AI-led discovery is reshaping brand visibility, media spending and consumer trust, and what it means for the future of digital marketing.



Edited Excerpts:


Your report finds that nearly half of Asian consumers are undecided at the start of their purchase journey. How are LLMs influencing their choices?


Parul Bajaj: Almost a quarter of all purchase journeys in India are now influenced by AI search or LLMs. That number has grown rapidly because, until 18 months ago, there was hardly any LLM search. Now, LLMs are among the top three sources of discovery for Indian consumers. After consulting LLMs, the brand consideration set expands. Consumers start considering brands they had not thought of before. So, whether a journey begins with a brand in mind or not, the consideration set expands.




Lauren Wiener: Globally, the figure is slightly lower, at about 14% outside India. Consumers are now influenced by about 15 touchpoints before making a purchase decision, up from five a few years ago. Marketers need to coordinate what LLMs read with what they put out on social media, TV, radio, billboards and other channels, so the message is unified.




India has traditionally been a Google-first market for product discovery. Are LLMs beginning to take meaningful share from traditional search?


Parul Bajaj: Social media is the single largest source of discovery for Indian consumers. Almost 30% of Indians look at social media as a key touchpoint when discovering what they want to buy. This is not just a high-income, metro phenomenon. Tier-two and tier-three cities, as well as lower-income households, also place high importance on social media in their discovery process.




India now has around 2.5 million monetised creators, and almost half are micro- and nano-creators. They have high engagement and targeted niche audiences, so their recommendations feel contextualised and personalised. There are parallels with LLMs. Consumers trust them partly because they get personalised recommendations that do not feel sponsored by a brand. There are a lot of similarities with India’s creator economy.




What does the rise of AI-generated recommendations mean for how brands compete for consumer attention?




Parul Bajaj: There are two things to consider: reach, or how often a touchpoint occurs in a shopper’s journey, and influence, or how much it affects the final brand choice. In India, LLMs feature in about a quarter of purchase journeys, but in 65% of those cases, shoppers say the LLM influenced their final brand choice. The only channel that comes close is social media.




Brands have traditionally allocated budgets based on reach, but they now need to consider reach and influence together. A channel with lower reach but high influence still needs to be optimised for. Brands also need the right content for LLMs to read.




Lauren Wiener: You need to be loved by humans and read by agents. To be read by agents, the approach differs from building brand love with people. Brands need to be present where LLMs draw information. Globally, key sources include Reddit, YouTube, LinkedIn, other digital spaces and a brand’s own website.




Content also needs to be structured differently. Charts, clear pricing and availability, and precise descriptions help because agents need quantifiable information to evaluate a purchase. Marketers need to build capabilities around Answer Engine Optimisation (AEO) and Generative Engine Optimisation (GEO), reverse-engineer the prompts consumers are using, and create natural-language content that responds to those prompts.




Does the rise of LLM search change brand-building, or does it simply shift discovery from search engines and social platforms to AI assistants?




Lauren Wiener: It's both. Google is itself a search engine and an AI discovery point. Compared with a few years ago, Google now offers AI-assisted answers and interactive conversations. The world has become more complex for CMOs: they need to be discoverable across 15 touch points rather than five a few years ago and maintain a consistent brand narrative across them. Information also needs to be written differently for agents.




Parul Bajaj: Historically, content was made for human consumption. Now, it also has to be readable by agents, which requires different skills. This affects the capabilities marketing leaders need to build in their teams and the capabilities agencies must bring. Brand-building budgets have traditionally been weighted more towards reach than influence. That allocation mechanism needs to change because a lower-reach channel such as AI search may have much higher influence than higher-reach channels.




Measurement and attribution also need to change. Digital media made it relatively easy to calculate return on ad spend for each dollar spent. With AI, there is no straightforward measure yet for how much is gained by optimising content for AI. We need to rethink measurement and attribution across touchpoints.




How should marketers rethink media-budget allocation as LLMs become more influential in purchase decisions?




Lauren Wiener: Historically, media budgets have followed eyeballs, although with a lag. Logically, advertising dollars should move towards LLMs as audiences grow. The shift will partly depend on how many ad slots platforms offer and which platforms decide to prioritise advertising. To the extent that they do, we will see more advertising dollars going to LLMs over the next few years.




If brands begin influencing AI-generated recommendations, could that undermine consumer trust?




Lauren Wiener: We have seen a version of this on platforms such as Reddit, where brands have tried to influence communities and faced backlash from people who feel they are being manipulated. It is a delicate balance: brands need to contribute authentically and provide balanced, useful content rather than propaganda. They should join the conversation, listen first, and respond helpfully to genuine questions instead of simply pushing brand messages.




Parul Bajaj: We need to understand how consumer trust changes as LLMs are monetised through advertising. Consumers currently trust LLMs partly because their recommendations feel objective, contextualised and personalised, rather than like marketing backed by brand spending. If ads are introduced into agents, we will need to reassess how consumers respond and how much they trust those agents. That remains to be seen.




The shape of budgets will change, not just the amount allocated to AI advertising. Brands will need to invest in content for humans and agents. This cannot be a one-time exercise. It needs to be an always-on process of checking what agents say about a brand and continuously updating and optimising content so agents can read it.




Could large brands have an advantage over smaller or emerging brands because more digital information about them is available online?


Parul Bajaj: That was our initial hypothesis: AI might narrow choices to a handful of big brands. But in more than half of AI-assisted journeys, consumers began considering brands they would not have found on their own. Those are not necessarily big brands. AI search may be creating a more level playing field.




The first step is to find out what agents say about your brand, then ensure the content is in the right form and format for them to read. In the past, discovery was about who outspent whom. Now, smaller brands can also put the right content in place.




Lauren Wiener: Globally, we have seen AI-native challenger brands gain share rapidly with LLMs because they are structured to follow this playbook. Historically, big brands benefited from reach built through television advertising that was expensive to produce and run. Today, anyone can create content at scale using AI and distribute it digitally, giving smaller brands a countervailing advantage. Our current hypothesis is that this will create a more level playing field rather than one that favours big brands.




How might agentic commerce develop in India, and how will its trajectory differ from China’s?




Parul Bajaj: China is one market where we have seen a more integrated agentic-commerce layer. Chat, social, commerce and payments are integrated within the same ecosystem there. India is different: chat, social and commerce involve different players, while payments through UPI are unified and democratised.




When we speak to marketing leaders in India, agentic commerce is one of their top priorities. Consumers are asking for it, and brands are planning for it, but the ecosystem is fragmented. It is likely to happen, though the form it takes may differ from China. My sense is that this will play out over the next 18 to 24 months.




Lauren Wiener: China was ahead in live streaming, influencer product lists and shoppable links. It is exciting to see India follow so quickly. The rest of the world is further behind in agentic commerce and shoppability, so there is a lot to learn. Across markets, creators are gaining influence over consumer opinion, with direct shopping links following as a lagging development.




How can marketers measure the impact of LLM visibility on sales when recommendations may not generate conventional clicks?




Lauren Wiener: There are leading and lagging indicators. As a leading indicator, marketers should track their share of mentions relative to competitors for the queries that matter most to them. For lagging indicators, it remains to be seen how quickly LLM feedback can be incorporated into traditional measurement models such as marketing mix modelling (MMM). The goal is to connect LLM mentions and readability to sales metrics.




What are the biggest risks for brands over the next two to three years? Is it becoming invisible to AI search, being inaccurately represented, or losing credibility?




Lauren Wiener: These risks are closely related. We are entering a world in which brands need to think of reach and influence as one connected journey and organise themselves accordingly. The priority is to ensure that the brand narrative is represented consistently. Traditionally, chief marketing officers have focused largely on controlling paid media. Now, they need to pay attention to what is being said across paid and owned media, as well as other touchpoints, including investor call scripts for publicly listed companies. AI agents will read information from all these sources, without necessarily prioritising them in a particular order. Brands therefore need to manage their narratives holistically across every conceivable touchpoint. Rather than thinking in terms of losing control, brands need to learn how to mediate conversations wherever they take place.




What does it mean in practice for brands to create content that AI agents can read and interpret?


Lauren Wiener: Leading brands across the globe have already started thinking about how to create content for AI agents. They are studying the queries consumers make and reverse-engineering answers to ensure that their content is readable by agents. This trend has already begun and is becoming another component of the digital marketing ecosystem. Companies that recognise the shared responsibility of marketing, product and technology teams are beginning to designate a single owner for the unified brand narrative. Digital marketing has become a two-pronged game. Brands need to win on both the consumer and the agentic frontiers, and their organisational structures must support both.




What are the biggest changes Indian marketers should prepare for over the next two to three years?




Parul Bajaj: One important trend is personalisation, driven by AI and agentic AI. India is a complex market, with thousands of languages, diverse consumer demand spaces, different income segments and distinct city-level markets. This makes it particularly important to deliver the right message, at the right time, through the right touchpoint. When we spoke to marketing leaders earlier this year, personalisation emerged as one of their top two priorities in India.




A second trend is the changing approach to media-spend allocation. Marketing leaders at some of India’s largest consumer companies are questioning their traditional methods of allocating media budgets and pushing their agency partners to rethink those approaches. I expect to see a significant shift in media-spend allocation over the next 12 to 18 months.




The third trend is the narrowing of the discovery-to-purchase funnel. Historically, discovery and purchase, particularly in the digital world, have taken place at different touchpoints. That funnel is likely to collapse or at least become narrower. Companies will need to consider how their marketing, sales, commercial and pricing functions are structured to reflect this shorter journey.




What role will agentic marketing play in these changes?




Lauren Wiener: Agentic marketing is an imperative for driving growth in the age of AI. Without it, costs could increase and organisations could slow down. Our clients are looking at agentic tools that can accelerate content creation without adding costs. They are also exploring tools that enable the next best action, support personalisation at scale, and improve media allocation by synthesising the five to 15 touchpoints that now underpin the collapsing funnel.




Marketers will need to encode their brand intelligence and context, essentially the knowledge and judgement of their best marketers, into a language that AI agents can interpret. That also means incorporating relevant regulations and rules into a format that both agents and people can understand. This will help power the agentic tools needed to drive growth without incremental costs.