At our recent Everyday AI Conference in New York, Kurt Muehmel, former Chief Customer Officer at Dataiku (who will transition to a new position at the company later this year), was joined on stage by Debbie Reynolds, Vice President of Enterprise Data Solutions and Engineering at Pfizer, for a conversation about the AI journey en route at Pfizer. Starting around ten years ago, with focus points of democratization and breaking down silos, Pfizer jumped into a process that has developed into a wonderful example of applied data analytics, one that encapsulates the essence of Dataiku’s championed concept of Everyday AI. In this blog, hear the highlights from Kurt and Debbie’s conversation to learn more about what Everyday AI looks like in practice for an organization like Pfizer.
What's Going On at Pfizer
Structured around people, process, and technology, Pfizer kickstarted a data-driven mission to transform its business processes through and through. Starting with the foundational goal of full data democratization, the organization has turned attention to the power held by descriptive analytics for both internal and external audiences.
Wishing to be at the forefront of analytics in the field, Pfizer focuses on empowering its analytical community to take advantage of the latest advancements in technology with software such as Dataiku. Through enabling collaborative analytics, opening up the dialogue surrounding predictive capabilities, and pushing to be more prescriptive in analytical initiatives, a very interesting evolution of the data and analytics processes at Pfizer is well underway.
A Supercharged Mission
At Pfizer, the pursuit of becoming a differentiated, data-driven organization that leads the health care community, is supported from the top down. Pfizer’s CEO supercharged the data and analytics journey at Pfizer through a mission referred to as Bold Moves. Each Bold Move innately involves advanced data science and analytics. Let’s explore a couple of these Bold Moves which Debbie described, and see how they manifest in practice.
Unleashing the Power of People
This Bold Move, “Unleashing the Power of Our People,” is an internally focused goal concentrated on using information about Pfizer employees to better their overall experience at the company. Data helps Pfizer understand the development of their employees’ careers, reveals areas to focus on to manage attrition, and offers insights into becoming more predictive about career progress.
Delivering Best-in-Class Science
This externally facing Bold Move, “Delivering Best-in-Class Science,” is all about getting the most advanced drugs to market faster and improving clinical trial recruitment. Using data-driven processes, Pfizer investigates how to deliver messages and products to customers exactly when and how they want them.
Additionally, Pfizer is focused on optimizing the way in which products are created. In the presence of what Debbie refers to as “ruthless automation,” Pfizer is looking at how they can remove components of human intervention in manufacturing while simultaneously implementing ongoing monitoring to check for any defects and ensure product quality.
People, Process, & Technology
Returning to the root of this journey that Pfizer is on, let’s look at the three main aspects of the AI transformation. Every part of Pfizer's core strategy right now surrounds driving these processes and advancements that require data and, more importantly, analytics. The end game that Pfizer has their eyes on is operations and transactions with fully embedded data insights.
Technology is the innermost component of the journey. Pfizer leverages Dataiku to systemize their data processes and Snowflake to improve data liquidity. Cataloging is also a critical aspect of their process so, Alation is part of the tech stack too. By creating an integrated platform to generate insights following a martech strategy, Pfizer is able to generate real impact from their data all the way through customer interactions.
But, without people, the transformation centered around technology would have no traction. Since the beginning of their journey, Pfizer has focused on casting away fears and building confidence in their teams through upskilling. Analysts have a variety of skill levels and areas of expertise, and it’s important to build comfort with the model process across an organization’s diversity of profiles. Everyone, from model building to business transaction teams, should possess the ability to articulate the way data is used and understand how and why models function.
The process that brings together the technology and people at Pfizer is centered around industrialization. When you’re upskilling people, it’s important to understand that even those who are more technically skilled will need to learn the different functions of the new technologies being onboarded. So, the process must incorporate an adoption of a flexible mindset — an acceptance of unfamiliarity and a willingness to adapt to new procedures. To make a case for the new technologies to your internal audience, it’s critical to simplify and support the people learning the functions and techniques. This is a part of the process that Pfizer tends to consistently.
Leaning In
With people, process, and technology components built out, there are still ways that Pfizer is working to lean into the journey and continuously improve these existing parts as well as the emerging demands that come with the increasing complexity and size of operations and integrations.
Debbie mentions that understanding the role of MLOps is one of these emerging areas for Pfizer. Another future focus for Pfizer’s journey is rationalizing and harmonizing the total AI architecture so that it works seamlessly within the ecosystem of organizational objectives. It isn’t all that broad though, one highly valued part of the journey moving forward for Pfizer is actually going back to basics and addressing issues around data hygiene.
Defining Success on a Data-Driven Wavelength
In the talk, Kurt asks Debbie how the AI initiatives support the core corporate strategy that Pfizer has laid out. The answer was simple.
These initiatives are going to be the basis from which success is defined for Pfizer moving forward. The Bold Moves are accelerators setting Pfizer up for the foreseeable future as a leader in the pharmaceutical industry. Through these data-driven initiatives, Pfizer hopes to not only address known customer needs with the right medicine at the right time but also hone the ability to address unmet needs.
These Bold Moves are a huge undertaking that requires an inherent acceptance of the fact that the analytics process does not belong to a particular function of the organization but rather the company at large. You have to make sure that data scientists are working in a transparent way and providing actionable insights that can be directly applied all the way to the point of decision-making to extract real value. AI applications need to be embedded throughout the entire organization at every step along the way. This is what Pfizer is working towards right now — a systemized state of AI that Dataiku likes to refer to as Everyday AI.