Key takeaway
Dashboards describe what happened; commercial teams need to know what to do next. How Sanofi and Servier use unified data, talk-to-data AI and decision thresholds to act faster.
Short answer: Start with one trusted source of commercial data, then move from dashboards that describe the past to tools that predict what will happen and recommend what to do. AI helps most at the last step: letting anyone ask questions in plain language and turning agreed thresholds into suggested actions. But the hard part is not technology. It is getting dozens of markets to agree on one version of the truth and deciding in advance who acts when a number turns red.
Most pharma commercial teams are not short of data. They are short of decisions. Reports show sales, market share, call activity and email opens, but they describe what already happened. Meetings then discuss the numbers instead of acting on them.
Filena Azevedo, Global Oncology Asset Lead at Servier, framed the problem on our stage at the 4th Pharma Omnichannel HCP Engagement Conference in Barcelona in March 2026:
"Dashboards don't change behaviour. Decisions do."
Filena Azevedo, Servier, 4th Pharma Omnichannel HCP Engagement Conference, March 2026
She distinguishes three layers of data: descriptive (what happened), predictive (what will likely happen) and prescriptive (what to do). Only the last two, she argued, give commercial teams the agility that fast-moving markets such as oncology demand.
Azevedo used a launch example. A rapid first-line uptake is good news, but it can also trigger a 60% probability of supply constraints, a risk of reimbursement delay in major markets and uneven regional penetration. Each signal maps to a pre-agreed action and an owner: supply chain accelerates production, market access invests in real-world evidence, and commercial reallocates field focus.
Her three-step approach for teams:
Vikram Jai Singh, Product Manager leading commercial omnichannel transformation at Sanofi, described where his team started: CRM, email, web, events, medical and messaging data spread across more than 15 systems, with no common HCP identifier and 50+ markets running their own processes. His diagnosis:
"This wasn't a technology problem. It was a change management challenge."
Vikram Jai Singh, Sanofi, 4th Pharma Omnichannel HCP Engagement Conference, March 2026
Sanofi followed a pilot, prove, scale path: 2 pilot markets, then 8 early adopters, then 30+ and finally 50+ markets. The unglamorous foundation was a common channel taxonomy, so that a "call", an "email" or an "event" meant the same thing everywhere. The turning point came when pilot markets saw every HCP interaction in one place for the first time. "Skeptics became champions overnight," he recalled in his presentation.
Once the data was unified, Sanofi added a generative AI "talk-to-data" layer. A user can ask, for example, which oncology HCPs engaged with a webinar but did not respond to the follow-up, or why cardiology engagement dropped last month, and get the analysis back without writing a query. Singh summed it up: "No SQL. No BI training. Just ask."
The value differs by persona. Field reps get a full HCP history before a visit; omnichannel teams get cross-channel attribution; leadership gets a consistent global view. Sanofi's next steps include predictive next-best-action, automated insight generation and, from 2027, agentic orchestration of campaigns.
The market is moving the same way. Salesforce reports field representatives spending up to 70% of their time on admin tasks, and both major life sciences CRM vendors now position their platforms around AI agents. Azevedo's advice for teams that are not there yet was pragmatic: "You don't need AI to speed-up action in your dashboards. But if you can, embed an AI agent yesterday."
Header photo: ThisisEngineering on Unsplash.
Descriptive analytics shows what happened, such as sales or market share. Predictive analytics estimates what will likely happen, such as a forecast or risk probability. Prescriptive analytics recommends what to do about it, such as reallocating field resources.
It is a generative AI interface that lets users ask questions about commercial data in plain language and receive analysis, without writing database queries or building reports. Sanofi uses it on top of a unified HCP engagement data foundation.
Because engagement data sits in many systems with different data models, no common HCP identifier and different channel definitions across markets. Sanofi's experience shows that fixing it is as much about change management as technology.
Not necessarily. Clear thresholds, pre-agreed playbooks and named owners already speed up decisions. AI agents add speed and scale on top of that foundation.
The Pharma Commercial Innovation Summit, organised by UBIQ Events, brings together pharma commercial leaders to discuss launch excellence, market access, omnichannel engagement and AI in commercial.
