SPEAKERS:
· Sihame Bahri, Global Omnichannel Lead, CSL
· Dr. Kishan Rees, Physician / Senior Director, Global Video Content & Digital Media Strategy – Patient & Scientific Engagement, Bayer
· Dr. Saima Sharif, Consultant Obstetrician & Gynaecologist, NHS
· Ed Hammerton, Head of Omnichannel Experience Design, Syneos Health (Moderator)
KEY TAKEAWAYS:
· LLM traffic to pharma websites reportedly trebled after ChatGPT Health launched, creating an unmeasured content audience
· Practicing clinicians notice pharma's educational withdrawal and consider it actively harmful to clinical knowledge
· Content designed for AI search can reach more HCPs — but not through pharma websites alone. Partnerships with trusted platforms and medical influencers are key.
· Platform-issued trust credentials are becoming a parallel credentialing system pharma neither built nor controls
· E-consent infrastructure gaps and field force repositioning determine whether omnichannel ambitions become operational reality
When OpenAI launched ChatGPT Health, something unexpected happened to pharma website traffic. "The traffic to pharmaceutical website companies trebled, quadrupled, went through the roof," Dr. Kishan Rees noted at Pharma 2026, "because it's taking that as trusted training data and source material. It's maybe B2LLM, we are putting stuff into the world for a whole host of companies and proxies to be the messenger of our content." The traffic spike pharma's analytics teams recorded may represent an audience the industry never planned for and cannot engage in dialogue: LLMs harvesting clinical content as authoritative source material, then redistributing it through AI-powered tools at a scale no direct HCP visit count captures.
That inversion set the frame for everything that followed. Ed Hammerton moderated the discussion and asked how peer-to-peer medical information sharing is changing and what the implications are for pharma. The panel arrived at a sharper question: why is pharma pulling back from engagement precisely as two forces, one artificial and one deeply human, are arguing for the opposite?
The Retreat Clinicians Can See
The digital transformation thesis that drove pharma's withdrawal from educational sponsorship and face-to-face engagement rested on a hypothesis: owned digital channels would deliver equivalent depth at lower cost. The evidence from this panel suggests they haven't, and the people best positioned to assess the damage are the clinicians on the receiving end.
Dr. Saima Sharif's testimony was direct. "I have seen since I was a very junior doctor a decrease in face-to-face interaction with pharma drug reps and educational events," she said. "One of the things I think should be revived is organisations like yourselves coming to those educational events and sponsoring them. Because education is key to get new information to health clinicians. I'm just surprised it's declining." This is a consultant obstetrician describing a trend she finds both visible and counterproductive, not a satisfaction survey data point, but a clinical professional naming a gap in how new evidence reaches practice.
Owned portals underperform, email open rates have plateaued, and the relational depth that face-to-face engagement generated hasn't been replicated in any digital format. The retreat was executed against a hypothesis the evidence hasn't vindicated. The space vacated doesn't stay empty.
Kishan placed this in terms that go beyond channel economics. "Stop optimizing for channels you can control, it is such a tiny, tiny, tiny proportion of the online conversation. Start earning our way onto channels that clinicians actually use," he argued. "Go where the bees are buzzing. If we don't, you look at the misinformation, we are letting patients down." The framing is deliberate: pharma's retreat from earned, clinician-used platforms is a failure of public health responsibility, not just a commercial miscalculation.
The voices filling the gap are not operating under pharma's regulatory frameworks. "Clinical content creators, digital opinion leaders and rising stars are filling the gaps left by Public Health England when it comes to the public health communication space," Kishan noted. "For life sciences not to engage with those people is almost a massive open goal, missed opportunity." These creators reach the audiences pharma has lost, without the clinical accountability pharma carries.
Hammerton synthesised the thread: "Almost like there's a returning importance of the face to face and also that kind of collective learning opportunities, combining your industry with those educational events where good solid research or educational information is being shared either online or face to face." The panel's convergence wasn't a call for digital retreat. It was a recognition that the human infrastructure pharma dismantled may be the scaffolding its digital strategy requires to function.
B2LLM: The Audience Pharma Didn't Know It Had
The "declining web traffic" narrative that drove pharma's disillusionment with owned digital content may be measuring the wrong thing entirely. LLMs are consuming pharma-owned content as training data at scale, creating a new content audience that inverts the business case for HCP portals and that standard analytics cannot see.
Hammerton crystallised the implication. "There is potentially a whole other audience. We talk about creating content not just for the healthcare professional, but also for the LLM, and structuring it in a way for them, and making the decision as to whether that's viable and whether that's valuable, is really critical." The concept Kishan labelled B2LLM reframes an underperforming channel as a misread one. Pharma content teams have been measuring direct HCP visits and finding them flat. The content sitting on those portals is simultaneously being ingested, summarised, and re-delivered to clinicians and patients through AI-powered tools that pharma neither controls nor tracks. The owned digital estate isn't failing. It's being redistributed through intermediaries the industry's measurement frameworks weren't designed to see.
The strategic implication cuts two ways. Pharma's investment in clinically precise, well-sourced portal content now has a second distribution channel with potentially far greater reach than direct traffic ever provided. But if that content is poorly structured, outdated, or clinically ambiguous, it is being amplified at scale by LLMs without any quality gate.
Sihame Bahri endorsed the B2LLM priority and immediately introduced the guardrail. "We need to optimise our portal for LLMs — that's the point if we want to still be relevant to our audience," she said. "But it's not the only thing, because there's this social dimension we should not overlook. HCPs and patients are the most social users because they are basically looking to solve a problem that is a disease." Replacing one tactical fixation, driving HCPs to owned portals, with another, optimising for machine audiences, would repeat the same structural error in a different direction.
Kishan described how Bayer is evolving its measurement framework to capture what engagement actually produces. Three dimensions matter: depth of engagement (did the viewer watch the full piece, return, share, and eventually teach the content to someone else?), behavioural change (the right patient receiving the right treatment, or the wrong patient not receiving it), and relationship durability over time. "That 90 seconds just helps us build awareness and trust and gets our foot in the door for those face-to-face interactions," he noted of short-form video. The inclusion of patient safety in a measurement framework, ensuring the wrong patient is not placed on a medication, elevates engagement metrics from commercial indicators to clinical ones. In a B2LLM world where content reaches patients through AI summaries without clinical mediation, that safety dimension becomes considerably more urgent.
Discover more on this topic at Pharma Commercial Data & Tech Europe 2026 (4-5 November, London) Europe's collaborative home for data and tech pioneers. Visit the website here.
Trust In Online Medical Information
In an information environment saturated with AI-generated content and unregulated health advice, the question of who signals credibility is being answered not by pharma and not by regulators, but by platforms and by audiences themselves.
Dr. Sharif pointed to a credentialing system already operating at scale. "YouTube put a trademark on healthcare professionals and the information they're providing to the public," she said. "When a patient comes and says this is what a doctor said on her webpage, she's a regulated doctor who's been providing advice. Those kinds of regulations on social media are very good and I think would be the future." Patients are using platform-issued trust markers to evaluate clinical content: a system pharma did not create and does not control, but which is actively shaping how HCPs' digital authority is perceived.
Hammerton drew out the structural significance. "There are now these hallmarks of trusted health information sources that social media platforms have introduced partly because there is so much disinformation and inaccuracy out there. It's important for clinicians to participate and to be active when they have a reputation there." Platform-level credentialing is emerging as a parallel verification system alongside traditional professional credentials, and clinicians who build presence in these ecosystems carry a form of verified authority that pharma's own channels cannot confer on them.
Bahri connected this to the evidence on what doesn't work. "I would see the rep or the MSL of the future as a social content moderator or content creator," she said. "Studies have shown that humans don't trust fake influencers. Industries ahead of pharma in social media, luxury, fashion, they tried AI influencers and pulled back. We as humans prefer to have a conversation, and we don't like perfection." The cross-industry evidence is instructive: synthetic engagement proxies fail because trust requires perceived human authenticity. Investment in human credibility infrastructure, field force as credible content creators, sustained clinician partnerships, educational presence, is not a nostalgic preference. It's what the data from adjacent industries indicates.
The Permission Layer Everything Else Requires
The debate between digital and human engagement, between B2LLM optimisation and earned social presence, shares a common dependency the panel addressed only obliquely. That dependency is the strategic constraint most likely to determine which organisations can act on any of this.
Dr. Sharif set the benchmark. "Pharma industries should create robust trusted connections with healthcare professionals on a long-term basis — not snapshots coming along here and there," she said. "We have a rep that comes from a medical device company every two to three months. If there are updates, she invites us to courses. That long-term connection is important." The bar isn't frequency. It's continuity. The medical device rep she describes is effective not because of visit cadence but because the relationship has enough history that an invitation carries weight.
Bahri identified the operational constraint most omnichannel strategies quietly assume away. She named e-consent collection as her single first priority, arguing that without an adequate base of authorised digital contacts, all downstream omnichannel strategies operate on an incomplete permission foundation. Without it, the B2LLM opportunity, the earned social strategy, and the field force repositioning all run on a contact base too narrow to generate meaningful signal.
The field force itself requires repositioning. "I would not start now and ask my teams to create new content," Bahri said. "What I would do is use the field force and make sure that HCPs understand how to work with AI, because then we can really show that we are a partner, not just another foot in the door company. Once the HCP understands the pitfalls of using AI, they can also pass it on to the patient." The rep as AI educator rather than product detailer is a role that creates durable value precisely because it serves the clinician's actual workflow problem.
Kishan offered the most concrete proof of what trust-through-ceded-control looks like in practice. The Heart and Kidney Care Alliance, he described, has "literally outsourced control of content and editorial and everything to 20 patient organizations around the world in the heart and kidney space. We don't want to own it or influence it. We want to provide the foundation, the platform to help the people that have been there, seen it, done it, worn the T-shirt." Relinquishing editorial control the industry usually insists on is pharma's most counterintuitive trust-building mechanism, and it most directly addresses the credibility deficit the panel spent the session diagnosing.
The organisations that will struggle are those treating the B2LLM opportunity and the human engagement imperative as sequential priorities, assuming they can fix the machine-readable content estate first and then return to the relational infrastructure. The panel's evidence suggests these aren't sequential at all. An LLM trained on thin, poorly structured portal content that no clinician trusts doesn't become more influential because it's machine-optimised. The credibility that makes pharma content worth amplifying, by LLMs, by clinical content creators, by platform-credentialed HCPs, originates in the same human trust infrastructure the industry has been quietly dismantling. Pharma that grasps this will treat the field force, the educational sponsorship, and the portal content strategy as a single investment. Pharma that doesn't will find its content amplified at scale, and its influence diminished in proportion.
To get you highlights of Pharma 2026 faster, we are using generative AI technology to summarise the transcripts of the sessions. If you have any feedback about the summary, please contact lucy.fisher@thomsonreuters.com.
Discover more on this topic at Pharma Commercial Data & Tech Europe 2026 (4-5 November, London) Europe’s collaborative home for data and tech pioneers. Visit the website here.
