Marketing conversations in boardrooms are evolving quickly, with the rise of generative AI pushing them into new territory. As consumers begin to shift their online behavior, many marketing leaders are now asking a new kind of question: what role should large language models (LLMs) play in our brand’s strategy?
In the past, product discovery was relatively straightforward. SEO, paid search, and social ads brought consumers into a well-mapped funnel. Today, however, a growing number of consumers are turning to large language models (LLMs) like GPT-4, Gemini, and Claude for product research and decision-making. These tools synthesize information across the internet into a single, trusted answer. For consumers, it is convenience. For brands, it represents a disruption that challenges long held assumptions about how and where customers find them.
Recent reports provide clear signs of this shift. Perplexity.ai reported more than 60 million monthly active users in 2025 (Perplexity, 2025). At the same time, product related Google searches declined by 14 percent year over year, according to Similarweb (2025). For younger consumers in particular, LLMs are becoming a starting point for research. They are not seen as a novelty, but as a useful tool. A Stanford University study found that more than 25 percent of Gen Z now start their purchase journey by consulting an AI assistant (Kleinberg et al., 2025). While a study by Bain & Company (2025) showed that 42% of LLM users ask for shopping recommendation.
Although this behavior is still emerging, the trend is clear. It does not necessarily signal the decline of search or traditional channels but it raises an important strategic question for companies that depend heavily on being discovered through those methods.
The question is not simply whether to adopt LLMs or not, but how to prepare for a world where they increasingly shape discovery.
Brands may find that their existing strategies don’t translate well into this new environment. LLMs do not display a list of links for users to choose from. Instead, they summarize, interpret, and recommend. Unless a brand is meaningfully present in the sources LLMs rely on, be it structured data, social media or reviews, it may not appear in the answer at all.
Even though this may feel urgent, it is better seen as a strategic turning point rather than a crisis. Some organizations are already taking steps to explore this space. For example, Harvard Business School uses virtual advisors in order to create a more controlled interaction, suggesting a model for how firms might deploy branded virtual advisors to influence LLM-mediated discovery (Bussgang, 2025).
Firms have also begun auditing how their brands appear across LLM platforms. They use structured prompt tests on tools such as ChatGPT, Gemini, or Perplexity to assess brand appearance (Luxury Presence, 2025; Writesonics, 2025).
Others are redesigning digital assets to optimize for AI answers. This includes updating FAQs and create schema rich product descriptions that align with how LLMs organize and retrieve information (Anderson Collaborative 2025). These initiatives might become the new norm in embedding brands into the new AI-first discovery ecosystem.
But these moves aren’t one-size-fits-all. For some companies, the priority may be to better understand how AI is influencing the customer journey before making major changes. Others may take a wait-and-see approach, choosing to observe the space evolve further before reworking media budgets or investing in new content frameworks. That too is a strategy as long as it is deliberate.
It is important that leadership teams begin to ask the right questions.
- How visible is our brand within LLMS-powered search and recommendation tools?
- Are we controlling our narrative, or relying on others to define it?
- Do we have the digital capabilities and data infrastructure needed to adapt if necessary?
- What type of conversion should we expect from LLM-powered discovery?
These are not technical questions alone. They are strategic in nature. They relate to brand positioning, customer trust, and long-term visibility. While these questions do not require immediate action, they do require thoughtful discussion.
Failing to engage with these developments may carry risks. Brands could lose visibility in key decision moments, fall behind more adaptable competitors, or miss opportunities to influence how they are perceived in new digital environments. Yet, over-investing too early, without clarity, can also be costly. Striking the balance requires a blend of experimentation, observation, and strategic alignment.
Ultimately, integrating LLMs into a marketing strategy is not simply about reacting to a trend. It is about understanding how the process of discovery is evolving and determining how, or if, your brand needs to respond. That answer will not be the same for every company but asking the question is no longer optional.
Still unsure where to start? Whether you're exploring how LLMs are reshaping discovery or rethinking your brand’s visibility in AI-powered environments, we’re here to help. At Merkle, we work with leadership teams to review, challenge, and evolve their digital strategies for the AI era. Get in touch with us here.
Author
Louise Sjödin
Senior Strategy Consultant, Merkle Sweden
