Background
AstraZeneca BeNeLux engages with healthcare professionals through many different channels, face to face visits, web, email, online and social media. All of these interactions generate a large amount of customer data. Until now, AstraZeneca relied mainly on fixed, rule based logic to make sense of that data. The team wanted to explore whether a different approach, powered by AI, could uncover patterns and profiles that static rules simply couldn’t capture.
Challenge
The core question was how to segment customers better, to turn scattered interaction data into a clearer, more useful picture. AstraZeneca is strong in building fixed rules, but rule based logic has its limits: it doesn’t easily surface the kind of hidden patterns AI can reveal. At the same time, there was genuine uncertainty going in. As Daan Vens, Innovation and Business Excellence Director at AstraZeneca BeNeLux, puts it: “I had no clue whether it would actually deliver anything, but we wanted to invest time in this with the team and find out.” Getting buy-in required a cross functional effort, bringing together the digital and IT side with marketing, commercial and sales teams, and pure commercial AI use cases were still relatively immature across the industry, which is why AstraZeneca chose to bring in an external partner rather than build this alone.
Solution
CROPLAND and AstraZeneca started with a proof of concept, a pilot phase built on a “learning by doing” mindset, with room for things that worked and things that didn’t. Rather than aiming for one big, distant end goal, the project was deliberately broken down into small, understandable steps, making the technology tangible for people without a technical background. Change management played a central role: AstraZeneca engaged its marketing community early, found the early adopters willing to experiment, and let momentum build from there. A key part of moving from pilot to business as usual was removing manual work wherever possible, automating data processes so the model’s output could become part of everyday operations, just like a sales team consults a daily sales dashboard. The ambition is a model that doesn’t just profile a customer, but actively recommends the next best action, for example, suggesting which channel and message to use to reach a specific healthcare professional.
Benefits
Results
AstraZeneca’s marketing teams are now using the resulting customer profiles directly in their campaigns, leading to more relevant customer interactions, which in turn generate more data to keep improving the model. As Daan Vens describes it, this creates a virtuous cycle that needed an initial push to get going, but that keeps reinforcing itself as it’s used. Looking ahead, Daan Vens sees AI eventually taking on a much larger share of a marketer’s day to day work, not by replacing sales people, but by generating faster, better insights and simplifying tasks that are currently still manual, freeing up capacity to focus where human expertise matters most.
“I think we’ve managed to profile our customers much better. Several of our marketing teams are genuinely using that information in their campaigns. And that’s how you learn, and generate more interactions, so the model can keep learning too,” says Daan Vens.
