SPEAKER:
Aurelio Arias, Director of Thought Leadership, IQVIA
KEY TAKEAWAYS:
· Only 20% of oncology physicians adopt new brands within two years of launch.
· Lifecycle compression occurs at launch, peak revenue and loss of exclusivity, with compounding impact.
· Annual commercial planning cycles have become a structural competitive liability, not a scheduling inconvenience.
· Behavioral, probabilistic physician models shift commercial intelligence from reaction to anticipation.
· Agentic AI compresses competitive intelligence workflows from days to seconds, but only where culture permits acting on probability rather than certainty.
On the surface, oncology is the specialty that commercial teams treat as a best-case adoption scenario: concentrated prescriber base, high clinical urgency, and sophisticated KOL infrastructure. The data says otherwise. Only 20% of oncology physicians prescribe a new brand consistently within two years of launch. As Aurelio Arias, Director of Thought Leadership, IQVIA, puts it, "the median is typically somewhere between three to four years… we actually see that tail of prescribing physicians sometimes seven, eight years after that brand has launched." If even oncology produces a seven-to-eight year adoption tail, the industry's general assumptions about launch trajectory are almost certainly worse than acknowledged.
The compounding problem sits on the other axis. "Brand plans and commercial strategy plans in general are typically created on an annual basis," Arias argued, "but the market is now moving much more quickly in a matter of days or weeks, months. And so we have a mismatch and the lack of being able to react is becoming ever more often an increasingly prevalent problem." When physician adoption is slow and the commercial window to capture value is shrinking simultaneously, every month of planning lag represents a larger share of total available revenue than it did a decade ago. The disadvantage is structural, not cyclical.
Three Vectors That Multiply, Not Add
Lifecycle compression is frequently discussed as a single pressure: generic entry, biosimilar competition, and pricing headwinds. The more accurate picture is three simultaneous forces operating at different points on the revenue curve, each making the others more damaging.
Arias described the front-end and peak-revenue dynamics directly: "Launches, for example, are being effectively delayed through stringent reimbursement requirements, such as head-to-head trials or longer trials to generate enough evidence… your peak sales are also being flattened a little bit through gross-to-net erosion, particularly in the US. And now we're starting to see, particularly through MFN, potential price caps as well." The European parallel reinforces the point. The EU's Biotech Act and Critical Medicines Act are reshaping incentive structures for rare disease and small-population launches in ways that will further delay reimbursement certainty in markets that have historically moved faster than the US.
At the other end of the curve, biosimilar maturation is eliminating what used to function as a financial buffer. Brands once enjoyed a long revenue tail toward the end of their lifecycle. Biosimilars are now "starting to compress a lot quicker," Arias observed. The long tail that historically absorbed the cost of a slow launch ramp is contracting precisely when companies need it most.
The compound effect is what most commercial planning fails to model. A delayed launch that was manageable when peak sales persisted for a decade becomes strategically crippling when the peak is simultaneously flattened and the tail amputated. Organizations treating reimbursement strategy, pricing defense, and loss-of-exclusivity planning as distinct functional workstreams are solving each problem in isolation while the integrated damage compounds.
The strategic prescription Arias offered challenges conventional launch sequencing directly: "Brands now need to focus really on the trajectory of increasing the launch trajectory much faster, be it through greater country coverage, so more markets much faster, or starting to stack indications rather than to sequence them. Bringing your medicine to as many patients as possible as quickly as possible because the future is less certain now." Indication sequencing was designed to extend exclusivity runway. When the runway is shrinking from both ends, the sequencing logic inverts.
Counting Prescriptions Versus Predicting Prescribers
The methodological argument Arias made is the most significant conceptual claim in the presentation. The historic approach, as he described it, involved "going out with curated data sets, being very sure and saying, okay, will we have a decision that's been made on A or B." The alternative is to take "the whole universe of physicians, whole universe of the prescribing decisions they make and medicines across time" and ask what patterns emerge in how physicians prescribe across a population. This is a genuine epistemological shift from asking whether a physician prescribed to asking when and why, and what that timing predicts about future behavior.
The dataset underpinning this approach establishes the foundation. "We have one and a half million prescribers in the US, billions of actual prescriptions that have been prescribed, and over 2,000 launches. That's the universe. That's our Launch Adoption Index," Arias described. Behavioral patterns at this scale become statistically stable in ways that curated, query-specific datasets cannot achieve. The signal-to-noise ratio that makes prediction viable only emerges from full-universe observation.
The segmentation findings carry immediate operational implications. [Arias] noted that "early adopters are typically principal investigators, they're heads of a department, they're linked to institutions, they're comfortable and they've been experimenting with these drugs for a while. Late adopters are typically junior physicians. So they wait for guidelines to be updated, they take a look at the data as a whole, they wait for key opinion leaders to provide input before they have the confidence to go out and prescribe." If late adopters are structurally guideline-dependent, then the timeline for guideline publication is a commercial strategy lever, not merely a medical affairs milestone. Companies that invest in accelerating guideline integration may compress the adoption curve more effectively than those that increase promotional frequency against already-resistant segments.
What Arias calls "a paradigm shift" in predicting adoption also implies something his presentation leaves implicit: launch underperformance can now be diagnosed in the first months, not in retrospective analysis years later. The gap between predicting adoption and changing it, however, deserves scrutiny. Knowing that a community oncologist will likely wait four years before prescribing is diagnostically valuable. Knowing what specific intervention compresses that to eighteen months is transformative. The Launch Adoption Index addresses the diagnostic; the therapeutic, what commercial actions actually accelerate late-adopter conversion at scale, remains the harder problem. Organizations that deploy predictive adoption models without redesigning the interventions those models should trigger will find themselves with better intelligence and unchanged outcomes.
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.
Agentic AI Eliminates Coordination, Not Just Calculation
The time-savings benchmarks Arias cited span different therapeutic areas and different analytical task types: competitive landscape analysis for diabetic macular edema saved over a day; a recent launch review for ulcerative colitis saved two days; cost-effectiveness analysis of promotional channels for oncologists saved another two days. "That might take several days, might take a week," he noted, "which now it can be produced in a matter of seconds." These benchmarks are self-reported against internal baselines rather than independently validated, a caveat that matters for calibrating expectations. What the breadth of applications does suggest is that the platform's value proposition is horizontal across analytical functions rather than deep in any single query type.
The underplayed implication is organizational rather than analytical. In most commercial operations, the bottleneck isn't the analysis itself, it's the queue of requests, the handoffs between data teams and brand teams, and the reconciliation of outputs pulled from systems that weren't designed to communicate with each other. A single natural-language interface that runs competitive landscape analysis, launch benchmarking, and promotional ROI assessment against integrated data sources doesn't just replace analyst hours. It eliminates the inter-team coordination overhead that routinely converts a two-day analytical task into a two-week decision process. That compression is where the strategic value concentrates.
The Organizational Bottleneck Neither Layer Solves
Arias closes with a four-part architecture that is internally coherent: behavioral science informing decision models, probabilistic rather than deterministic indicators, leading signals to prepare for what's coming, and woven together "agentic AI to be able to react and action on those signals much faster." The logic holds. Predictive adoption intelligence and agentic execution address both the anticipation and the response problems that annual planning cycles cannot.
What neither layer provides is the organizational willingness to act on probability rather than certainty. Pharma commercial culture is built on deterministic confidence: validated segments, confirmed market access, and proven message effectiveness. The shift to probabilistic, behavioral intelligence asks brand teams to make consequential resource decisions, reallocating promotional investment, redirecting field force effort, adjusting launch sequencing, based on likelihood rather than proof. Companies that deploy these tools into organizations that still require certainty before action will produce the worst possible outcome: they'll see the signal faster and still wait too long to respond.
The diagnostic is straightforward. If a brand director, presented with a probabilistic signal that a competitor's launch is accelerating adoption among their late-adopter segment, cannot reallocate resources within days without convening a steering committee, the technology investment has solved the wrong bottleneck. The annual planning cycle isn't just a scheduling artifact; it reflects a deeper organizational norm about what constitutes sufficient evidence for action. That norm is what lifecycle compression is actually pressuring. The brands that close the gap between signal and action won't be the ones with the best data or the fastest models. They'll be the ones that changed what counts as good enough reason to move.
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.
