Transforming Manufacturing at Pfizer: The Hard Part Was Not the Technology
Introduction
In pharmaceutical manufacturing, every production run is governed by a batch record, a detailed, step-by-step document that provides instructions and then captures what was done, by whom, and whether each parameter fell within approved limits. For decades, these records have been paperbased: thousands of pages per batch, filled out by hand, and archived to satisfy regulators. If any potential quality issue arose, release of the batch may have been held while quality assurance analysts reviewed it to ensure that the product could be released.
For major pharmaceutical players, like Pfizer, the value of implementing a modern manufacturing execution system (MES) with electronic batch records (EBRs) is obvious: Replace paper records and manual documentation with electronic systems that can monitor and capture every step as it occurs, while supporting real-time analytics and improving efficiency.
For nearly two decades, Pfizer, along with the rest of the industry, struggled to capture that value. Industry attempts mostly produced what practitioners dismissively called “paper on glass”: electronic forms that replicated traditional batch records on a screen while shop floor equipment remained disconnected and data sat in local silos. Nothing came close to a platform that could provide AI-powered, real-time analytics.
By 2025, Pfizer had broken through. What had been a perpetually behind-schedule transformation program was doubling the rate at which it converted manufacturing sites year over year, at a fraction of the original cost estimates. Manufacturing cycle times decreased and quality investigations dropped. And the concept of digitized shop floor became the foundation for AI-driven process innovations, insights, and optimization.
Pfizer’s success came from solving the problems that had blocked progress for almost 20 years: building trust with shop floor operators; forging partnerships between digital teams and manufacturing operations; creating repeatable technology deployment processes.
This case study traces the obstacles Pfizer encountered during its early attempts to implement EBRs, the management practices that overcame them, and the returns that followed. The lesson at its center extends well beyond pharmaceuticals: The most consequential barrier to AI-powered transformation is rarely the artificial intelligence. It is the unglamorous work of getting the underlying systems, data, and human relationships right.
Building the MES Backbone: Review-by-Exception as the Gateway to AI
By Susan S. Garfield and Srihari Rangarajan
Pfizer’s ambition to build a predictive, data-driven manufacturing ecosystem required more than deploying technology. It demanded deep coordination across site operations, central digital teams, and transformation stakeholders. The objective was to move beyond “paper-on-glass” limitations toward an integrated Manufacturing Execution System (MES) and Electronic Batch Record (EBR) that connect shop floor execution with enterprise decision-making in real time.
A defining element of the approach was establishing a truly collaborative delivery model across Pfizer’s global network of manufacturing sites and digital stakeholders. Rather than providing sites with centrally designed solutions, plant operators, quality teams, and digital leaders jointly defined requirements, shaped solutions, and drove adoption. This coordination was critical in overcoming siloed processes, limited trust, and fragmented data environments.
The global digital organization implemented a modular MES/EBR architecture that balanced standardization with site-specific flexibility. Rigorous, site-level diagnostics, cross-functional workshops, and continuous engagement aligned deployments to operational priorities such as cycle time reduction, QA disposition efficiency, and throughput optimization. This integrated approach enabled stronger ownership at the site level while maintaining enterprisewide consistency in data and processes.
Coordination across stakeholders was further reinforced through disciplined governance, phased delivery, and transparent performance tracking. Deployments were broken into value-driven increments, enabling early wins and building momentum across sites. By embedding teams alongside operations and fostering ongoing collaboration between digital and manufacturing, the transformation evolved from a technology program into an operations-led initiative with strong stakeholder alignment.
The impact of this coordinated approach was measurable. Transitioning to MES and EBR enabled real-time data capture, review by exception, and streamlined quality processes. Manufacturing cycle times were reduced by over 11%, and quality outcomes significantly improved.
Ultimately, Pfizer’s transformation highlights core fundamentals of scaling digital manufacturing: tight alignment across technology, operations, and people. By orchestrating coordination across site teams and digital stakeholders, MES and EBR become more than systems — they form the backbone for sustained operational excellence and a foundation for future capabilities, such as AI-driven optimization and predictive manufacturing.
Susan S. Garfield, DrPH, is an accomplished strategist, consultant, governing board member, and thought leader who leads major accounts for the global EY organization and serves as director of the Center for Health Outcomes Optimization.
Srihari Rangarajan leads supply chain services for life sciences clients across the Americas. He has more than 20 years’ experience developing innovative supply chain solutions for clients in the pharmaceutical, medical device, and consumer product industries.
Challenges in Pfizer’s Manufacturing Environment
Pfizer’s manufacturing network — more than 30 sites worldwide — presented obstacles that would test even the most sophisticated digital transformation programs. The company’s diverse product portfolio created vastly different needs for EBR deployment. Work on sites ranged from high-volume solid oral dose manufacturing to aseptic processing, biotech operations producing monoclonal antibodies, and clinical-scale production. Biotech operations were particularly complex, involving the cultivation of biological organisms through multiple growth stages — from tiny vials to progressively larger vessels up to 12,000-liter bioreactors spanning three floors. Even the smallest event within kilometers of stainless-steel piping could destroy multiple batches in sequence and require weeks to remediate. No single software design would work across this diversity.
Compounding this was the company’s legacy of mergers and acquisitions. At one point, Pfizer operated more than 60 plants before consolidating. Each site had different automation layers and control systems. As John Sourke, vice president of global biotech operations, explains, some sites had a single integrated system running the entire plant, while in others, “you buy equipment from 20 different vendors. They all have their own automation systems, all have their own PLCs [programmable logic controllers], all have their own SCADAs [supervisory control and data acquisition systems], and none of them talk to each other.” Various digitization efforts had produced what Joe Cozzolino, senior director, digital manufacturing and digital global supply, calls “islands of digital” that couldn’t support integrated batch records or analytics. What should have been standardized deployments became custom integration projects at nearly every site.
Regulatory considerations added a further layer of difficulty. Each production line’s operations must be approved by regulators, and modifying a manufacturing process to fit a standardized system could require costly refilings. This made the intuitively appealing concept of one-size-fits-all standardization essentially impossible. Meanwhile, the company’s legacy automation software operated with what digital leaders characterized as “1980s and 1990s era” technology that made data access difficult and expensive.
Challenges in the Original Transformation Approach
Pfizer’s earlier MES strategy involving the launch of four major plants at the same time quickly proved overly ambitious. The program consumed a significant portion of the organization’s digital portfolio investment, which was unsustainable.
Also, the original vision — building solutions centrally and deploying them to sites — positioned manufacturing plants as recipients of technology rather than participants in its development. Governance was murky: The boundaries between central digital responsibilities and site responsibilities were poorly defined. Sourke explains, “The accountabilities between what was being driven centrally versus what was being driven locally, and who was pulling and who was pushing, weren’t exactly clear.”
Moreover, the digital team lacked manufacturing credibility. Staffed primarily by systems experts with limited grounding in pharmaceutical operations, the team struggled to earn trust at sites. Engaging shop-floor subject matter experts was especially difficult when the efficiency gains could be perceived as possibly leading to head count reductions. Schedule attainment ran at only 60%, meaning go-live targets were missed 40% of the time.
The situation came to a head at one of Pfizer’s largest sites. Sourke says, “It was a big site that was used to really delivering on everything they do.” The site’s leaders, frustrated, demanded the program restart with a new technology vendor, saying, “We’re not getting the response we need. We’re spending money and not making progress. This is not how we do things.” While the divergence from the standard approach at Puurs ultimately gave the site a sense of ownership over the solution, it represented a significant failure to maintain the standardization that technology leaders had envisioned.
Rebuilding the Foundation: A New Approach
The transformation leadership team knew that trying harder with the same approach would yield the same results. They needed fundamental changes to how they thought about, and executed, the MES/EBR deployment. Sourke notes, “I was just frustrated with the lack of progress we were making. Every time we arrived at a new site, it was difficult. Too difficult.”
So Sourke offered to take charge of the transformation process. He gathered senior executive support to rethink the transformation program and appointed a key manufacturing leader to partner with IT leaders and consultants. Having someone on the team with the authority to drive change — someone who knew plants inside out — complemented the IT team’s technical knowledge. IT, for its part, hired Cozzolino, who had extensive manufacturing operations experience, to manage the delivery program.
From implementation to enablement. The first breakthrough came from reframing the mission. Instead of implementing MES, this new team would focus on enabling MES capabilities. This shift in wording represented a fundamental shift in mindset — positioning the team as partners helping facilities achieve operational goals, with MES as a means to an end. The new approach centered on a common architectural framework that could be systematically customized for each site’s context, maintaining consistency where it mattered while allowing necessary variation.
Strategic site selection. With resources constrained and credibility on the line, the team also developed a decision framework balancing potential value against implementation complexity. Cozzolino explains, “One of the things that we found was we’d jump into these sites without a clear understanding of their value drivers. If we could cover 80% of revenue in a plant at 50% of the cost of full transformation, we could make better progress.” The new framework helped to build credibility through a series of wins rather than bogging down in lower-volume, but more complex, production lines at each site.
Rigorous diagnosis before action. The team committed to thorough upfront preparation — sometimes lasting three months — for each site before starting technical work. This proactive step involved understanding not just the technical landscape but operational realities: how work actually flowed, where pain points were, what informal work-arounds existed, and the organization’s capacity for change. The diagnosis period culminated in a collaborative implementation road map with specific business outcomes, not just technical deliverables. It also engaged colleages at the site early, so they could help with the implementation rather than feel it was imposed upon them.
Showing value at every step. The team broke implementations into smaller phases, each delivering measurable results. They established clear metrics, tracked progress through dashboards and regular reviews, and celebrated wins. These changes maintained momentum, supported continued investment, and provided an early warning when adjustments were needed.
Managing the human element. Cross-functional teams combined manufacturing operations experts with IT specialists. Rather than parachuting in with a predetermined solution, the teams worked alongside site colleagues, understanding their challenges, respecting their expertise, and incorporating their input. As Vijay Patel, vice president, manufacturing technology and solutions, puts it, “Having the humility to say, ‘We’re not the experts; you are the experts. We’re the experts in the technology and solutions that can help you do your job. But ultimately, you need to be in the driver’s seat.’” Each team invested heavily in training and support, staying engaged after going live to troubleshoot issues and refine the systems after seeing how they performed in the real world.
Building organizational capability to change. Beyond individual implementations, the transformation team built repeatable processes to accelerate delivery. They developed comprehensive implementation templates — from diagnostic worksheets to standard configuration patterns — that provided starting points for customization rather than dictating rigid solutions. Recognizing that not every site needed full MES complexity, they created an “MES-lite” offering for smaller or simpler operations. They also applied AI tools to analyze diagnostic data, identify patterns across sites, and assist with documentation. Each implementation became faster, more predictable, and more successful than the last.
Results
The proof of any transformation lies in outcomes. While Pfizer’s MES effort is still ongoing, the results validate the revised approach. Yearly EBR deployment has increased by more than 50%. The team’s go-live schedules have turned out to be more than 80% accurate, compared to 40% accuracy two years ago. The cost of implementing an EBR at a site has come down by 80% as a result of unlocking scale and making MES economically viable. And the time involved has been compressed due to templated approaches, MES-lite options, and accumulated organizational capability.
Operationally, sites with MES deployed have achieved better “right first time” quality, according to the Six Sigma principles that Pfizer follows. On average, quality assurance disposition lead time is down and full production cycle time is faster. Plants now have access to reliable, real-time data that enables better decision-making at every level — from line supervisors optimizing daily production to corporate leaders benchmarking performance across the network. Early skepticism gave way to demand: Plant managers under constant pressure to improve performance are now requesting MES implementations.
The Enabling Power of Getting the Basics Right
Pfizer’s transformation is not a story about AI, advanced analytics, or cutting-edge technology. It is a story about what those capabilities require before they can deliver value and how the work of building that foundation is itself transformative.
For most of two decades, the company’s opportunity to digitize the manufacturing execution process had two main obstacles. The technology itself wasn’t ready — the team was dealing with rigid systems and unconnected machines that required extensive manual activity and paper forms. And organizations weren’t ready — they lacked the governance, credibility, and deployment discipline to execute even when better tools emerged. When integrated EBR capabilities matured enough to connect directly with automation systems and capture real-time process data, deployment challenges remained. The barriers were organizational: misaligned priorities, transformation teams that lacked manufacturing knowledge, a model of one-size-fits-all software that was technically attractive but didn’t fit on-the-ground realities, and a deployment model that treated sites as targets rather than partners. Pfizer’s breakthrough came not from better technology but from a fundamentally different approach to engaging the people and processes that technology depends on. According to Doug McHugh, senior vice president of digital, “This isn’t a digital project now. It’s an operations project that has a digital component to it.”
What makes this case instructive beyond pharmaceuticals is the sequence. The MES foundation, painstakingly built through trust, disciplined processes, and hard trade-offs, is what made everything that followed possible: “golden batch” optimization, AI-driven process control, real-time quality disposition, and a network-wide data infrastructure that turns every connected site into a source of insight for every other. None of those capabilities could have been layered onto paper records or disconnected islands of digital. The foundation was not a preliminary step to be rushed through along the way to more exciting work. It was the work.
The foundation is not just technological. In its transformation program, Pfizer learned that, as former IBM CEO Lou Gestner said, “The soft stuff is the hard stuff.” An architectural vision that made sense to technology teams — creating a one-size-fits-all package that could be easily maintained from headquarters — ran counter to the realities that the company’s manufacturing leaders lived. Rather than mislabeling the pushback as an unwillingness to accept standards, the transformation team needed to change its approach. Shifting the mindset from MES implementation to MES enablement empowered the transformation team to increase trust, build a shared understanding, and collaborate with site leaders to cocreate a system that worked for both headquarters and sites. Cozzolino describes the difference: “At the beginning, nobody wanted to talk to us, but now it’s dramatically changed. It’s really now a partnership. At the sites, we’re getting their top people. They own this now.”
For organizations pursuing AI-powered transformation, Pfizer’s experience offers a pointed lesson. The capabilities that generate the most excitement — machine learning, generative AI, and autonomous agents — are only as powerful as the systems, data, and human relationships underneath them. Companies that rush to build exciting new features on a shaky foundation of a core technology, systems, data, and relationships — that treat modernization as a cost to be minimized rather than a capability to be built — will find themselves perpetually unable to extract value from the tools they most want to use. The unglamorous work is not the obstacle to transformation. It is the transformation.
