The rules of pharma discoverability have changed

For decades, pharma brands built their market presence through two separate channels: direct-to-consumer advertising that sent patients to their doctors and physician education that pointed prescribers toward clinical guidelines and reference databases.

Now, AI is changing all of that. Large language models (LLMs) are already part of everyday clinical practice – more than four in five physicians report using AI in their work, more than double the rate from just three years ago. Patients are following suit: nearly a third of U.S. adults say they turn to AI chatbots for health information or advice.

This shift doesn't just introduce a new channel for pharma brands. It redefines what it means for them to be found in the first place. While search algorithms reward the right keywords and the right bids, AI generates responses based on the sources it was trained on and has learned to treat as credible.

Most pharma brands already have the evidence they need to influence that association. But before they can act on it, they need to understand where they stand.

Right now, many brands are focused on the wrong thing. Rather than looking for ways to show up in models without a plan, brands should ask what the model thinks about it right now and whether that perception is reaching physicians and patients the right way.

Only once you have that baseline can you make decisions about what content to create, where to publish it and how to structure it in a way that LLMs can actually absorb it. As AI continues to become more commonplace, the brands that show up will be the ones LLMs have learnt to trust.

Shaping the question

But showing up in an LLM response assumes the right question was asked. Physicians may inadvertently ask clinical copilots the wrong questions and patients seeking health information from a chatbot may leave out important details. In either case, the LLM will simply answer what's asked rather than volunteer what's missing.

That gap is where pharma's real opportunity lies. The goal isn’t just showing up in the answer but becoming a source LLMs already trust, so that a brand's information is there before the question is even posed.

Turning insights into action

Gaining that understanding requires more than making content discoverable in common places. Optimising purely for how a specific LLM works today is a short-term play. What sources it pulls from and how it ranks credibility can shift every time a model is retrained. The brands that build long-term visibility focus on content quality and authority signals that hold up across updates, not just the mechanics of today's platforms.

Through our AI platform, WPP Open, we’ve developed two tools to help pharma keep pace. Agent testing allows us to query AI models directly to see how they perceive a brand, condition or treatment. Synthetic audience adds another valuable layer. It allows us to create digital representations of physicians and patients we can query to understand the questions they ask, measure how AI shapes their perceptions and test how different brand strategies perform.

However, there's a constraint for pharma to consider. Brands may want to be authoritative sources for LLMs, but they can't – and shouldn't – share sensitive patient information or commercial data to get there. WPP's Open Intelligence offers a safer path. Brands can connect their first-party data with external sources to generate richer intelligence and train AI models, all without that data ever being moved or exposed.

Armed with all of that information, pharma can make smarter decisions about what comes next. As AI becomes the new front door for medical information, the content that gets cited looks nothing like a traditional brand website. Instead, it lives in peer-reviewed journals and owned properties structured like research hubs, which are free from the marketing language that LLMs filter as noise. None of this approach works, however, without continuously tracking how AI represents your brand and then adjusting as the signals change.

Ultimately, discoverability is an intelligence challenge before it is a media challenge. The pharma brands that get this right won't just show up in a response. They'll be there when it matters most – before the question is even asked, when a physician is deciding on a treatment or a patient is searching for answers.