SPEAKER:
Fleur Gill, VP Sales & Partnerships – Global, Doceree
KEY TAKEAWAYS:
· Nearly half of prescription volume sits beyond consistent field force reach today
· Coverage gaps are architectural constraints, not headcount or effort problems
· Pharma's compliance infrastructure converts AI scale from liability into strategic asset
· AI agents extend breadth; human reps retain depth in high-value physician relationships
· The competitive axis is shifting from field force size to coverage architecture design
"Nearly half the opportunity lives in part of the territory, or even part of the model, that you can't consistently reach."
That figure, offered by Fleur Gill at Pharma 2026, should stop any commercial leader cold. Territory design in pharma has been refined over decades. Decile-based targeting, segmentation models, hybrid coverage structures, the industry has applied serious analytical rigour to the question of where to send its people. The result of all that optimisation is a model that works exceptionally well for the top tier of prescribers and leaves a substantial portion of the physician universe systematically underserved. This is not accidental neglect. It is the logical output of an architecture built to maximise depth, not breadth.
Gill, whose company Doceree markets an AI engagement platform for healthcare professionals, has a commercial interest in the diagnosis she offers. That context belongs on the table. It does not, however, make the diagnosis wrong. The structural argument she advances is separable from the product she sells, and it is the structural argument that commercial leadership teams should sit with. "It's not a staffing issue, and it's not a sales issue," she argued. "It's an architectural issue. And if you think about architecture and architectural problems, they're not usually solved by adding more weight to a structure that's already under strain."
That reframe changes the conversation entirely.
The Ceiling That Headcount Cannot Raise
Pharma's standard response to coverage shortfalls has been role proliferation. MSLs layered onto traditional reps. Key account managers added above them. Digital engagement specialists running alongside. Each addition expands capacity within the existing model without altering the model's fundamental constraint: physician time. A rep can only call on so many physicians per day. A physician can only receive so many visits. The ceiling is set on the demand side, not the supply side, which means adding headcount on the supply side is, at some point, a category error.
Gill frames this directly. "When a model hits a ceiling, the efforts stop being the answer. So you can ask people to do more and you can ask your teams to stretch further, but that ceiling in terms of coverage will just stay where it is." The implication is not that field force effort lacks value. It is that effort applied inside a structurally constrained model produces diminishing returns regardless of its quality. Underperforming coverage metrics are not, in the first instance, a management or motivation problem. They are a design problem.
The strategic reorientation that follows from this diagnosis is pointed. "The question is no longer how do we push harder inside those top deciles in our territories — it becomes how do we build breadth in the whole physician universe?" Intensity within existing reach has been optimised. The variable left unsolved is reach itself. This pivot from depth to breadth as the primary coverage challenge represents a genuine shift in commercial strategy logic, not a tactical adjustment.
The mechanism Gill proposes for extending that reach is AI-driven engagement: agents trained on brand and product content that can sustain compliant, personalised interaction across physician segments that human reps cannot cover economically. In her framing, "commercial reach and commercial presence is no longer tied just to human availability", a structural decoupling that, if it holds at scale, rewrites the economics of coverage design.
One caveat is worth noting here. The 'nearly half' figure draws on Gill's perspective as a practitioner rather than an independently verified source. The directional claim is plausible. Decile-based targeting has always concentrated effort on a fraction of the prescriber base, and industry analyses consistently show that mid- and lower-tier prescribers receive substantially less engagement than their aggregate prescription volume would warrant. But the specific proportion is asserted, not evidenced. Commercial teams building a redesign case internally will need to validate the gap against their own data before the architecture argument can do its full work.
When Governance Becomes the Growth Lever
"You do not want scale to come at the expense of control." Delivered in the context of AI deployment rather than compliance management, this statement inverts the usual framing. The standard narrative positions regulatory rigour as the price pharma pays for operating in a sensitive category — a drag on speed, a tax on agility. Gill's argument runs the other way: control is not what you sacrifice to achieve scale; it is what makes scale deployable at all.
The evidence for this comes from outside pharma as much as within it. Consumer-facing industries that deployed AI chatbots without governance infrastructure have generated brand damage through hallucination, off-message responses, and uncontrolled drift. The failures are well-documented. Pharma's MLR review process, claims substantiation requirements, adverse-event monitoring obligations, and role-based access controls were built over decades to govern what human representatives can say to physicians. That same apparatus, Gill argues, is the pre-existing framework that makes AI-driven physician engagement safe enough to deploy. The regulated environment that slows pharma down is also the condition that makes its AI adoption more defensible than AI adoption in less-governed industries.
The knowledge architecture she describes maps directly onto this compliance logic. An AI agent trained on "product information, efficacy data, pivotal trials, safety information" with the organisation controlling the content layer, is not a general-purpose language model turned loose on physician conversations. It is a constrained system operating within an auditable boundary. "It can be trained on the information around your products, your brands and the information ecosystem around them," Gill explained. "You're in charge of it. Importantly, it's MLR ready." Compliance is built into the architecture, not retrofitted after deployment.
The full governance stack she enumerates "MLR readiness, approval, claims, evidence, auditability, fair balance, role-based access and privacy" is typically discussed in pharma as a checklist of requirements. Gill reframes it as a competitive asset: "These are what turn scale from being a risk into an asset if done right."
That reframe carries a distributional implication the presentation does not address. The compliance-as-enabler argument implicitly advantages organisations with mature MLR infrastructure already in place. Large pharma companies with established review processes, governance teams, and content management systems are structurally better positioned to deploy AI agents safely than smaller biotechs or emerging companies that may lack these systems at scale. If AI-driven coverage extension becomes a meaningful commercial differentiator, the companies best placed to capture it are the ones that have already invested most heavily in the compliance apparatus that enables it. The "living sales organisation" may widen competitive asymmetry rather than democratise coverage reach. Organisations without that infrastructure should treat it as a prerequisite investment, not an afterthought to AI procurement.
There is also an organisational culture dimension. Companies whose MLR function operates as a gatekeeping process — sequential, adversarial, slow — will find the transition harder than companies that have evolved compliance into a design partner for commercial content. The technical capability and the governance mindset need to move together.
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.
Expansion, Not Displacement
"RepTwin is not a replacement model, it's an expansion model." This is the line designed to neutralise field force resistance, and it will face the predictable scepticism that accompanies any "this won't take your job" claim from a technology vendor. The scepticism is reasonable. It should not obscure the structural logic underneath the claim.
The human-depth / AI-breadth boundary Gill draws is the model's organising principle, and it rests on a meaningful distinction. Human reps operating in complex therapeutic areas — oncology, rare disease, immunology — are doing something that goes beyond information transfer. "The field force, MSLs, sales teams — they do something that's incredibly valuable and they create depth with healthcare professionals and doctors," she argued. "Their relationship remains essential because at the end of the day they've got judgment, trust and exchange taking place." These are not qualities that current AI systems replicate. They are the qualities that make high-value physician relationships commercially durable over time.
What AI agents offer is different in kind, not just in degree. Gill describes the capability as analogous to deploying "your best rep, your best MSL, that strongest pharmaceutical rep — it's just like that, but with the persistence and scale of software," capable of engaging "hundreds if not thousands of physicians at one time, not just dozens." The physician segments that sit beyond consistent human reach — mid-tier prescribers in underserved territories, specialists who decline in-person visits, early-career physicians building formulary habits — are precisely where persistent, compliant, always-on engagement could shift prescribing behaviour at the margin.
The depth-versus-breadth distinction is elegant in theory. The open question is whether it holds under operational pressure. As AI systems improve in conversational sophistication, the boundary between breadth engagement and relationship-building will become less fixed than the current model assumes. Organisations adopting this framework should plan for boundary management as an ongoing discipline, not a one-time design decision.
Coverage Design Is the New Commercial Differentiator
Gill's argument, followed to its logical conclusion, creates a new competitive axis in pharma commercialisation. For decades, share of voice correlated with field force size. The company with more reps covering more territory won more prescriber attention. That equation has been under pressure from access restrictions and digital channel proliferation for years, but the underlying logic of more people equals more reach has remained the default planning assumption. If the coverage ceiling is structural, and AI agents can address the physician universe that no human field force can touch consistently, then the differentiator shifts from how many people you deploy to how intelligently you design the system in which those people operate.
Companies that treat AI agents as a cost-reduction mechanism with fewer reps doing the same job cheaper, will misread the opportunity entirely. The value is not in substituting cheaper capacity for expensive capacity within the existing reach boundary. The value is in extending reach beyond that boundary to physician segments that currently generate prescription volume without receiving proportionate commercial engagement. That is a growth argument, not an efficiency argument, and it requires a different investment thesis.
"The future model is not human and it's not AI, it's actually a coordinated system in which both do what they do best: humans go deep, and something like RepTwin goes nice and wide, and together, when you think of coverage and that ceiling we keep running into, it's going to remove the coverage ceiling." The word "coordinated" carries the most weight in that sentence. This is not a technology procurement decision. It is a commercial architecture decision requiring orchestration capabilities, data integration, content governance, engagement sequencing, performance attribution, that most organisations are not yet built to manage.
Gill's three-part summary names what that architecture replaces: "from scheduled reach to intelligent presence, from a field force model to a commercial system, from fixed human capacity to integrated coverage design." The diagnostic question for any commercial leadership team is not whether AI agents can extend coverage. The technology trajectory makes that question largely settled. The question is whether the organisation's compliance infrastructure, data systems, and commercial culture are mature enough to make that coverage trustworthy. For organisations that can answer yes, the coverage ceiling may be genuinely removable. For the rest, the ceiling has simply acquired a new name.
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.
