We’ve entered the era of artificial intelligence (AI) and there’s no going back. About half of all Americans now use large language models (LLMs) such as ChatGPT, Gemini, Claude, and Copilot, and that’s just the tip of the iceberg. Behind the scenes, an estimated 750 million apps use LLMs to power everything from chatbots to product descriptions to personalized recommendations.
The shift to AI is disruptive in countless ways. One area we track closely is how AI adoption and use impact digital marketing and the ability of biopharma companies to get their content in front of relevant new audiences. A lot of the discussion on this topic focuses on the potential of generative engine optimization (GEO) (aka AI optimization [AIO]), as search engine optimization (SEO) provides diminishing returns. It’s worth exploring, however, as detailed below, the opportunities for GEO are not guaranteed as the technology continues to evolve.
Diminishing Returns on Search Engine Optimization
Let’s start with what we know to be true: Opportunities to drive organic traffic to a website via search engine optimization are declining. LLMs are expected to overtake Google as the primary tool for search in 2028. This shift should be factored into long-term search optimization strategies, but it’s worth noting that it’s still in the early stages. In 2025, Google remains dominant, processing over 15 billion queries a day and maintaining just under 90 percent of the global market share for search. And given Google generates around $54 billion in ad revenue every quarter, its executives are unlikely to cede market share without a fight. Indeed, we’ve already seen the introduction of AI into Google’s search engine platform, creating more of a hybrid product. And this is changing how the technology is used.
If we dive a little deeper into the data, we see individuals are using search engines like Google differently. There’s a notable rise in what are known as “zero-click” searches. As of December 2024, a Bain & Company survey found approximately 8 out of every 10 search engine users now rely on AI summaries at least 40 percent of the time. In other words, they enter their query and never click through to a web page or result. They’re satisfied with Google’s AI answer and either exit or move onto another search (often prompted through features such as “people also ask”). Zero-click searches have big implications for SEO, along with paid search engine marketing tactics. Even if the search volumes remain strong, they often don’t result in users clicking through to the website. This signals a major shift in the traditional “buyer’s journey” that marketers have mapped for many years.
The declining market share of Google and the rise of zero-click searches will increasingly limit the number of visitors that are channeled through search engines to relevant websites. To replace this lost traffic, attention has shifted to LLM models and how they pull their information. If users are favoring ChatGPT or are satisfied with an AI summary from Google, how can we ensure that our content features there?
The Opportunities and Limitations of GEO
First, a gut check on the limitations of GEO. This may be obvious, but it’s worth calling out: Having your company’s content incorporated into AI answers isn’t the same as driving traffic to your website via SEO. In the past, a lot of effort went into creating educational content that would rank well in search results. A company developing non-opioid pain relievers, for example, might share educational content about the biology of pain. If optimized, this page might appear in search results, drawing users onto a company website where they can then learn about a new therapeutic approach.
Today, AI may still reference that content, but it is presented as an overview within the same interface (be it ChatGPT or Gemini). Surveys have found less than 20% of users always or usually click through to the sources supporting AI overviews. You’ve helped individuals learn, but they haven’t learnt about you. So, it’s important to evaluate your goals and what success will look like. If your goal is to deliver key messages on a topic––for example, highlighting an unmet need––that may be supported by a smart GEO strategy. However, if your goal is site traffic and conversions, GEO (and increasingly SEO) is unlikely to deliver ROI.
On that note, what is the return on investment, and can you really track it? Search engine marketing has historically provided a wealth of data. Referral traffic can be traced from search queries all the way through to form-fills on a company website. But if your web content is now appearing in AI results and no actions are taken, it’s hard to determine who saw what and whether it influenced their behavior in any way. This inability to track engagement and influence is thus another major challenge.
While these are significant considerations, there are still opportunities to be meaningfully included in LLM answers. If this is your goal, there are a few important nuggets to keep in mind. The first is that the fundamental principles of SEO still hold true. There are some areas where SEO and GEO diverge slightly. For example, we recommend adding “summary” sections at the top of key web pages to help LLMs synthesize the information. But while there are other minor shifts, for the most part, SEO rules generally apply and for good reason. Google has spent decades refining its algorithms to prioritize high-quality, highly relevant content for its users. Large language models follow a similar playbook, rewarding well-structured content that demonstrate attributes like E-A-T-T (experience, expertise, authoritativeness, and trust).
Given the overlaps, we don’t recommend dramatically changing your approach to digital content development.At CG Life, our philosophy for SEO has always been to focus on quality content and user experience and we recommend a similar anchor for GEO. Focusing on the customer experience and providing genuine value aligns our strategy with the ultimate goal of the technology companies, be it Google or OpenAI. Their business depends on a positive customer experience, so it’s a safe bet as the industry evolves. Trying to game the algorithms with SEO or GEO hacks is short-sighted, as the algorithms quickly update and evolve to close loopholes and penalize sites that employed them.
Spanners in the Works? Closed-Domain Systems
Digital marketers––particularly SEO/GEO practitioners––love to sound confident even when there are factors and algorithms that are categorically unknowable (I say this with affection!). It’s how the industry has evolved, but now more than ever, we need to be humble about what we don’t know and cannot anticipate. The rate of progress is too fast, and new models and systems can completely undercut an established approach.
One pertinent example in our industry is OpenEvidence, an AI platform for health practitioners. OpenEvidence is what is known as a closed-domain system. It is designed to support “high-stakes clinical decisions,” so it draws exclusively from peer-reviewed papers in reputable journals. By limiting the scope of its data inputs, it hopes to minimize the influence of unreliable information (think: Wikipedia, social media, and personal blogs), as well as hallucinations.
While OpenEvidence is just three years old, it has impressive momentum. In just a few short years, the company claims to have engaged more than 40 percent of physicians in the United States. The team recently closed another large financing round, valuing the company at $3.5 billion.
If OpenEvidence continues to grow and evolve, it will crowd out the use of search engines and general LLMs for professional medical searches (arguably a good thing). It may become the portal for how HCPs gather and interpret new information, and they will be routed to snippets of published papers not websites. For biopharma companies, this could have major implications for HCP marketing, including branded and unbranded HCP websites that would previously have been optimized for search. Here, a simple shift from SEO to GEO falls short as OpenEvidence exclusively pulls from academic journals. More emphasis may need to be placed on alternative channels, including paid media placements (you can advertise via OpenEvidence directly), or more subtly, through your scientific publication strategy.
OpenEvidence is specifically for HCPs, but other curated closed-domain platforms may emerge to serve information to other stakeholders, including investors and patients. We can all agree the internet encompasses the best and the worst of humanity. In that light, it makes sense to limit where information is sourced.
Where Do We Go from Here?
The rate of adoption of new technologies is intensifying. Transformations happen seemingly overnight. In this new landscape, biopharma companies need to stay nimble with tactics like SEO and GEO, as we can’t fully anticipate what new technologies will dominate in the future.
What does that look like in practice? Our recommendation is, first and foremost, to lean into an omnichannel approach, with multiple channels activated to engage a given audience. This provides a steady stream of data and allows you to pivot your strategy if new technologies arise, such as closed-domain systems that make SEO/GEO difficult.
Secondly, to embrace current opportunities for SEO/GEO, consider focusing 90 percent of your content marketing energy on relevance and quality. The final 10 percent can be dedicated to structuring the content favorably for search engines, artificial intelligence, and––crucially––web accessibility (we’re big advocates for the latter!).
We’re always happy to chat if you’re interested to learn more. We have plenty of tactics for optimizing and polishing content, but it’s not wise to overinvest in any given direction when the rate of technological change is so rapid. And beware the agency that tells you it’s the next big thing with no caution for the future!
