SPEAKER:
Daniela Seixas, CEO, Tonic Easy Medical
KEY TAKEAWAYS:
· Tier 3 passive physicians achieved 58.4% email open rates when triggered by real-time AI query intent
· 70% of European physicians now use medical AI as their primary clinical information source
· Italian GPs are already querying triple-therapy retatrutide before any regulatory market entry
· Physician AI query volume detected Mounjaro overtaking Wegovy in mind-share ahead of prescribing data
· Physician concern has inverted from AI dehumanizing medicine to personal over-reliance on it
The Ground Has Already Shifted
The channel migration is complete. "70% of physicians already use medical AIs as their primary source of information," Daniela Seixas told attendees at Pharma 2026, citing survey data from 770 European medical doctors collected in November 2025, figures she noted are consistent with published literature across Western markets. Pharma's push-based engagement infrastructure was engineered for a physician attention economy that no longer exists. Segmented email cadences, rep call schedules, portal-based detailing: each was built for physicians who sought information through channels pharma controlled. The question is not whether physicians have moved to AI copilots. It is what those copilots reveal about physician intent in the moments that matter most.
The answer, at least in one asthma case, is striking. "If you do it in real time, that is a doctor asks a question in our AI copilot that includes asthma or a related keyword and you send this email 15 minutes after the interaction, then it's the Tier 3s who respond better, the ones who are passive for asthma that don't respond very much to email, which with a 58.4% open rate plus 13 percentage points from the mass email." The hardest-to-reach physicians responded best, not despite being passive, but precisely because the outreach arrived at the moment their clinical attention was already engaged. This is not an optimization of existing engagement models. It is a structural inversion of the targeting logic that underpins most commercial segmentation.
The 15-Minute Window Demands Different Infrastructure
The asthma open rate is compelling. What it demands from pharma organizations is considerably more challenging than the headline suggests.
Seixas's platform is not a research panel or an opt-in survey population. "The engagement in our AI copilot is growing exponentially by physicians since May 2024. And we are serving a community of roughly 200,000 medical doctors in Europe, across France, Italy, Spain, and Portugal, which is our test market." That scale matters because it transforms physician query behavior from anecdote into a continuously instrumented behavioral dataset, one generated inside clinical workflows rather than retrospectively reported through surveys.
The 15-minute trigger model operates on a logic that current pharma CRM architecture was not designed to support. Standard commercial systems are optimized for batch segmentation: identify a cohort, craft a message, schedule a send. Intent-triggered engagement requires event-driven technical infrastructure, sub-hour content deployment pipelines, and crucially, medical-legal review processes that can approve communications at a speed batch workflows have never needed to achieve. Most pharma organizations, even those with sophisticated digital capabilities, are not structured to close that compliance gap quickly. The 15-minute window is simultaneously a product design challenge, a regulatory affairs problem, and a vendor relationship negotiation. All three must be solved in parallel before the signal becomes actionable.
The competitive context sharpens the urgency. Traditional medical information platforms are ceding ground as AI copilots grow, with legacy players "losing relevance as they didn't adapt fast," while Seixas's platform's power users are growing "at three digits." The differentiation, she argues, lies in clinical proximity: her platform is "really vertical," with AI models operating in local languages and medicine databases restricted to products commercialized in each specific market, making it structurally closer to the point of clinical decision-making than global platforms generating generic scientific summaries.
For commercial teams evaluating this opportunity, the infrastructure gap between signal availability and operational readiness is the real strategic constraint. Pharma companies that wait until the internal plumbing is built before engaging these platforms will find the engagement window has been claimed by competitors with fewer compliance constraints on their speed.
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.
Queries as Intelligence, Not Just Engagement
The commercial case for physician AI platforms extends well beyond triggered communications. Longitudinal query data functions as a leading indicator, earlier, cheaper, and more behaviorally proximate than claims databases or primary market research.
"If for example we look at when Wegovy was launched we see tirzepatide moving up even though Novo was in an important moment of the market trying to push semaglutide for obesity... Mounjaro in terms of brand was able to be on top of mind of the doctors ahead of Wegovy." This signal was visible in real time, inside the platform's query database, before any prescribing data could surface the shift. Pharma competitive intelligence teams typically operate on 60–90-day lags from claims audits or sales data. A query-based signal compresses that window to hours. The strategic value is not incremental, it changes the decision-making timeline for commercial response.
The obesity longitudinal case extends this further. Tracking physician questions across the same platform from 2023 to 2026 traces a legible evolution in clinical understanding: from foundational GLP-1 education, through molecule comparison and adverse event queries, to questions about therapies not yet on the market. Seixas notes that Italian GPs are "already asking about the triple therapy retatrutide," ahead of any regulatory entry. For launch teams, this is pre-launch demand sensing at a granularity that traditional market research cannot match. It also identifies which markets are generating early clinical curiosity versus which remain in earlier stages of therapeutic adoption, a geography-specific intelligence layer with direct implications for field force deployment.
Perhaps the most operationally provocative implication: the "emerging indications that doctors are asking about semaglutide and tirzepatide are quite different, probably reflecting the work of the MSLs." If physician AI query patterns can detect downstream effects of MSL activity through a channel the field force does not directly control, this opens an entirely new category of field force effectiveness measurement, one that bypasses self-reported call outcomes and captures actual physician information-seeking behavior.
One qualification is warranted. The competitive brand data and longitudinal question evolution are generated and analyzed by the same entity selling access to them. The directional findings are behaviorally grounded, but independent validation of these intelligence signals against prescribing outcomes would substantially strengthen the commercial case before organizations build budget commitments around them.
Governance Is the Physician Trust Infrastructure
Regulatory architecture, in this context, is not a compliance cost. It is the mechanism through which physician trust is earned and sustained at scale.
Seixas describes a governance stack built for clinical credibility: "We already own ISO 42001. Actually having an audit at the office right now to make sure that we are tackling governance of AI at a company level, not just at the product level. And our drugs model for the physicians is medical device Class IIa." Under MDR, Class IIa classification subjects the medicines model to conformity assessments and post-market surveillance obligations that generic AI platforms do not face. In a market where 70% of physicians have migrated to AI as their primary information source, the platforms that earn clinical trust will command the prescriber attention pharma currently pays to acquire through traditional channels. The governance investment is a credibility infrastructure play, not a box-checking exercise.
The complication is behavioral. "Doctors shifted from thinking that AI would make medicine less human as per our data to, at a later data point from November last year answered by almost more than 700 doctors, being worried about their own over-reliance on AI." This inversion matters analytically. If physicians begin to distrust their own AI-mediated reasoning, the intent signals generated by their queries become harder to interpret. Does a query reflect genuine clinical need at the moment of patient care, or habitual platform dependency? Pharma AI engagement strategies are not currently designed to distinguish between these two behavioral states, and the difference has direct implications for how triggered outreach is calibrated and measured.
Seixas's December 2025 expansion to pharmacists surfaced a categorically different information-seeking profile: predominantly brand name, dose, price, and practical calculation queries, including real-time BMI calculations. The contrast with physician query patterns is material. Pharmacists are not evaluating therapeutic positioning; they are resolving transactional clinical questions at the point of dispensing. Content strategy by channel cannot be a scaled-down version of the same message. It requires a fundamentally different brief.
The Intermediary Paradox
The strategic paradox embedded in Seixas's argument is one she has no structural incentive to name. Physician AI copilots are displacing the platforms, publishers, and representative interactions through which pharma historically shaped the clinical narrative. The same platforms now offer pharma access to the richest HCP intent data available. But accepting this access means becoming a data customer of the intermediary that displaced pharma's information authority in the first place. The deeper the integration, the deeper the dependency, and the more leverage the platform accumulates over the terms of that relationship.
This is not an argument against engagement. The asthma trigger data, the obesity intelligence timeline, and the competitive brand signal all point to a genuinely differentiated intelligence layer. But pharma organizations entering these partnerships should do so with clear-eyed recognition of the power asymmetry: the platform owns the physician relationship, the behavioral data infrastructure, and the engagement window. Pharma rents access to all three.
Seixas acknowledges the discipline this demands: "As new data sources proliferate, selectivity is critical. We believe in the quality of our data and you should look for data sources with quality." The selection criteria matter as much as the decision to engage. And the pipeline through which raw signals become commercial actions carries its own risks: "Beyond the quality of the data is of course your AI pipeline because this can also bring biases to your insights or can also help you articulate through agents and do real-time marketing automations like I showed you for example in the asthma product growth case."
Before evaluating any physician AI data partnership, commercial leaders should answer three questions that sit entirely inside their own organizations: Does our CRM architecture support event-driven, sub-hour triggered communications? Can our medical-legal review process approve content at the speed the engagement window demands? And are we building internal analytical capacity to interrogate externally sourced intent signals, or outsourcing intelligence judgment to the same platform that is also selling us access to it? The signal is real. The infrastructure to act on it is the work that remains.
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.
