From ‘Google it’ to ‘LLM it’

For years, our online journeys have started in the same way: type something into a search bar, scroll, click, compare, repeat.

While many brands have been optimising for position zero, consumers have moved on to position none, instead asking LLMs (Large Language Models), like ChatGPT, Claude, and Perplexity for recommendations. And the question is, where does your brand sit in that conversation?

This shift in the way consumers search is especially relevant during busy shopping periods like Christmas and Cyber Weekend, when consumers are looking for guidance on what to buy. LLMs provide a way to explore options conversationally, helping people get results faster and more efficiently than traditional search alone.

It’s not that search is disappearing. It’s the way we search is evolving. LLMs change the interaction from a list of blue links to a conversational back-and-forth. You don’t browse; you ask and refine. You get to the point faster. And once you’ve experienced that, it’s easy to understand why it’s becoming part of more people’s everyday digital habits.

Search is more fragmented

Nonetheless, traditional search engines still matter, they’re just now part of a wider discovery mix. People might look something up on Google, check it with an LLM, and cross-reference social content. It’s messier, more fluid, and more context-driven than ever.

For brands, visibility isn’t defined only by keywords or page rankings. What matters is the quality, clarity, and consistency of the information provided.

If product details, messaging, or claims differ between a brand’s website, social channels, retailer listings, or press coverage, LLMs may struggle to determine which version is correct. Clear, aligned, and credible information gives models something reliable to work with, helps brands appear accurately and helps direct consumers.

How consumers interact with LLMs

This also depends on which LLM people use. Each LLM has its own ‘personality’, shaped by its safety rules, ethics, and the sources it relies on. That means two people asking the same question won’t necessarily get the same recommendations.

Some consumers gravitate toward models designed with stronger ethical guardrails, like Claude from Anthropic. Others default to whatever is built into the platforms they already use, such as Llama in WhatsApp or Grok in X. And because each LLM draws from different sources and applies its own reasoning, the landscape of recommendations varies by category.

LLMs are often perceived as more neutral and authoritative than traditional search, particularly when they explain reasoning or cite sources.

What works for brands

Brands can use this understanding to improve how they show up in LLMs by focusing on:

• Ensuring messaging is consistent across channels

• Understanding which sources the model considers authoritative in their category

This varies significantly by sector. For example, according to Digitas’ Model Sight, in travel, we found that around 80 per cent of sources referenced by LLMs for hotel-related prompts came from Booking.com, meaning OTAs now play a major role in how hotels appear in recommendations.

Other categories are far more complex. In banking, for instance, models draw from a much wider and more fragmented set of sources. Here, alignment across websites, social channels, and press coverage becomes even more critical.

High-quality, credible, and consistent information strengthens visibility. Conversely, conflicting or inaccurate messages can dilute how a brand appears.

The impact on traffic and conversions

Taking these steps to improve how brands show up on LLMs isn’t just about visibility; it also affects outcomes. There are early signs that traffic to websites from traditional search may be declining. At the same time, referrals coming directly from LLMs often appear to be higher quality, with people more likely to convert a sale.

Rather than focusing purely on website visits, brands may need to rethink digital goals. Websites are increasingly acting as repositories of authoritative content that LLMs and agents can access, rather than just destinations for discovery.

Additionally, emerging standards like the Multiple Context Protocol (MCP) point toward a future where LLMs can draw directly on brand data and content for interactions.

Exploring tools to understand LLM visibility

To navigate this new landscape, tools like Model Sight are helping brands understand which sources LLMs rely on and how they are referenced. These insights guide marketing teams in keeping messaging aligned, credible, and optimised for AI-driven discovery.

The brands that will thrive are the ones that make it easy for models and consumers alike to understand who they are, what they stand for, and why they can be trusted. That means clear messaging about products and services, consistency across websites, social channels and press coverage, and content that comes from credible, authoritative sources.

In a world where LLMs are increasingly part of everyday decision-making, from holiday shopping to routine choices, these elements are no longer optional; they’re essential.

Leila Seith-Hassan is chief data officer at Digitas UK, and Caitriona Gallagher is strategy partner at Digitas UK.