Optimizing Biomanufacturing with Digital Tools
With digitalization gaining traction in biomanufacturing, The Pharma Navigator speaks to experts about how digital twins and advanced modeling are helping companies meet industry demands and what that means for the workforce.
Digitalization of biomanufacturing processes can provide companies with significant cost and time efficiencies, which can ultimately provide the end user with therapies at a quicker and more cost-effective rate. Many biomanufacturers have already made good inroads into investing in digital solutions, such as digital twins and advanced modeling, to help them meet the evolving demands of industry.
To find out more about how companies are leveraging digital twins to optimize biomanufacturing, the most beneficial applications of advanced modeling and simulation for biopharmaceutical production, and what workforce reskilling initiatives may be required, The Pharma Navigator spoke with a panel of experts. The panel comprised: Alexander Seyf, CEO and Co-founder, Autolomous; Minni Aswath, VP of Process Development and PD Downstream, Bionova Scientific; Dan Strange, CTO, Cellular Origins; Mike Tomasco, Senior Vice President, Chief Information Officer, FUJIFILM Biotechnologies; and Max Baumann, co-founder and partner, Treehill Partners.
Leveraging Digital Twins
TPN: How are digital twins being leveraged to simulate and optimize biomanufacturing processes?
Seyf (Autolomous): Digital twins are being used in biomanufacturing to simulate how processes behave in real time, allowing companies to test and adjust things virtually before making changes in the actual facility. This is especially useful in areas like scaling up from lab to production, where traditional trial-and-error methods can be slow, costly, and risky. By using digital twins, companies can explore different process conditions, like changing temperature or feed rates, and predict how these will impact yield or quality, without interrupting live production. For example, Sartorius has shown that digital twins help teams optimize processes much earlier by running virtual experiments, reducing the need for repeated lab work and shortening development timelines.
On the operational side, early adopters see stronger process consistency, better traceability, and fewer unplanned deviations, delivering clear reductions in cost and cycle time while improving control and quality across the board.
To get results like this across the entire industry with digital twins and advanced technologies, we first need to make sure every company, no matter its size, can reach an essential level of digital capability.
Aswath (Bionova): Digital twins are gaining traction in modeling bioprocesses, allowing virtual experimentation without jeopardizing live production. Bionova uses advanced simulations to optimize upstream parameters such as media composition, feeding strategies, and bioreactor conditions. Tangible benefits include faster process development cycles, early risk identification, and reduced time and cost for tech transfers. These models also help refine scale-up strategies, contributing to smoother transitions into GMP manufacturing.
Strange (Cellular Origins): Digital twins deliver substantial value when their insights directly guide operational decisions. In biomanufacturing, they're increasingly being used to model workflows and predict potential bottlenecks during facility design and initial deployment, ensuring that equipment, resources, and processes are optimally arranged from day one.
Moving forward, digital twins also hold significant promise for real-time optimization. Combined with flexible and modular manufacturing systems, digital twins could enable dynamic rerouting or reconfiguration in response to unexpected events or changing demands. Companies exploring these capabilities anticipate benefits including improved throughput, reduced downtime, lower failure rates, and accelerated production timelines, all contributing to faster, more reliable patient access.
Tomasco (FUJIFILM): Digital Twins provide an exciting opportunity within biopharmaceutical manufacturing. The real promise of this capability is your ability to simulate processes and outcomes without having to perform the experiments in real-life. It is possible to run through many variations of inputs and experiments with process parameters to optimize your process yield.
Once you have an optimized model of your process, the real value comes from monitoring your bioprocess in real-time and predicting outcomes. Those predictions can be used to inform the actions that you take to ensure your process stays on track for the best possible outcomes.
One example that is becoming more common place is the modeling of the Golden Batch for a specific product/process. This involved examining your past batch history to discover your optimum runs and the process characteristics that yielded those results. Process analytics teams use that historical data to model that scenario and multi-variate variations of what may impact that process and the types of interventions required to maintain the Golden Batch profile. When deployed on an Edge Computing device and run in real time, the outputs of the predictive model can be shared with the operator of the process. Within the past 12 months Generative artificial intelligence (GenAI) has been added into this capability with prior training on all of the process documentation and critical process parameters/interventions that have been successful in the past.
GenAI is then able to interpret the graphs and background information to suggest the next best step for the operator to maximize the output of the process. This human-in-the-loop use of advanced analytics and GenAI is helping biomanufacturers advance their digital maturity while creating real-world results with higher quality product outputs.
Baumann (Treehill Partners): In any new-build that is currently being proposed or conducted by a leading provider, all these tools are being used and there is fierce technology competition in play. Not all market participants in the manufacturing industry rely on technology leadership though. Which means, yes we see significant opportunity to make use of real-time, virtual replicas of production processes, that enable companies to simulate, monitor, and optimize every step in silico even before building the actual infrastructure, but not always this is ‘requested’. Where utilized today, this has led to reduced batch failure and more consistent product quality, probably resulting in faster time-to-market but n here is still small.
On the regulatory compliance side, the whole digitalization with enhanced monitoring and automated documentation has strongly increased productivity.
Impact of Advanced Modeling and Simulation
TPN: Beyond predictive maintenance, what specific applications of advanced modeling and simulation are proving most effective?
Tomasco (FUJIFILM): Advanced modeling and simulation, supported by AI and data analytics, are revolutionizing biopharmaceutical production by optimizing manufacturing conditions to achieve higher yields and faster production times. AI-powered algorithms enhance quality control through continuous data monitoring, ensuring compliance with safety standards and reducing batch failures. This improves reliability and accelerates production transitions, keeping companies like FUJIFILM Biotechnologies at the forefront of healthcare innovation.
Strange (Cellular Origins): Predictive maintenance is just the starting point. The real value of advanced modelling shows up when we use it to understand and respond to variability within the process as it happens. One major area is timing and resource orchestration. In a real facility, batches can start at different times, and delays can arise. If we can simulate these common scenarios ahead of time, we can plan how to sequence them through shared equipment more effectively, avoiding the kind of bottlenecks that kill productivity or lead to missed windows for critical steps.
Additionally, advanced modelling significantly enhances biological process optimization. For example, instead of relying solely on fixed schedules, real-time monitoring combined with historical data enables adjustments, such as determining the optimal time to harvest based on actual cell growth and viability. This results in improved cell expansion, higher-quality outputs, and fewer QC deviations.
The essential factor is translating rich, real-time process data into actionable insights. If a delay occurs, the system can dynamically adapt, reprioritizing batches and reallocating resources without compromising overall quality. Rather than seeking theoretical perfection, advanced modelling and simulation build agility, resilience, and readiness for confident scalability.
Seyf (Autolomous): Advanced modeling and simulation are making a significant impact in biomanufacturing by enabling real-time process control, more efficient scale-up, and better risk management. One key area is the use of soft sensors, models that infer hard-to-measure attributes like nutrient levels, metabolite build-up, or viral vector titer from real-time data inputs. These allow operators to adjust parameters dynamically, maintaining optimal conditions throughout production and reducing the risk of batch failure. In cell and gene therapy, where variability is particularly high, these adaptive control strategies are proving especially effective in improving consistency and yield.
On the scale-up front, simulation tools are helping teams model transitions from small-scale to commercial manufacturing, identifying potential bottlenecks early and avoiding costly surprises. Rather than relying on trial-and-error, manufacturers can predict process performance at scale, which not only speeds up tech transfer but also reduces the number of failed scale-up attempts. Collectively, these modeling strategies are streamlining production, enhancing quality, and enabling faster delivery of therapies to patients.
Baumann (Treehill Partners): Real-time process optimization, digital process control, and ‘what-if’ scenario analysis, all improve manufacturing yields and reduce failures in a prominent way today. Mechanistic and hybrid models enable dynamic process parameter adjustments, detect deviations early, and ensure optimal conditions during both upstream and downstream processing.
For the biotech industry, scale-up has always been a tricky thing being capital-intensive and often an afterthought after achieving clinical proof-of-concept. Over the coming decade, we expect scale-up simulations will allow also smaller manufacturers to accurately translate lab-scale results not only to clinical but also to commercial scale, minimizing costly surprises and delays as well as allowing them to get their assets as well as themselves as companies more comprehensively ready for moving into the commercial marketplace.
Aswath (Bionova): Advanced modeling and simulation are being applied beyond equipment monitoring to optimize yield, by modeling process parameters like temperature and pH in real time; reduce batch failure through dynamic control strategies informed by historical process data; and accelerate scale-up by simulating outcomes at pilot and commercial scale before physical implementation. Bionova leverages these tools within its Quality by Design (QbD) framework, improving consistency, reducing deviations, and enhancing overall productivity.
The Future Workforce
TPN: What workforce reskilling initiatives are required as the industry shifts towards highly automated and digitally integrated biomanufacturing environments?
Strange (Cellular Origins): The future workforce is a hybrid one. Operators will need to collaborate with robots. Engineers will need fluency in both bioprocessing and control systems. Quality teams will need to validate digital workflows, not just paper ones. We’ve built our system to be intuitive, modular, and accessible. Our interfaces are designed for process scientists, not just automation engineers. Our system architecture supports gradual adoption, letting teams build confidence incrementally. Most importantly, our platform augments the workforce — it doesn’t replace it. It lets people do more of what they’re best at, with less room for error and rework.
Baumann (Treehill Partners): No industrial workforce today can be called ‘skilled in AI’. Even the big techs’ teams keep learning every day. Skill is commonly focused on one’s own desktop. Rarely has the use of AI become a purposefully designed organizational process where human participants along the workflow leverage it in a coordinated, repeatable, scalable, trainable, auditable, way. Most people still use AI user interfaces to improve their own work output, rather than to improve the working process and organizational flows. We then look at AI based processes like the ones discussed above, and mostly these are seen as ‘products’ that are readily made and deployed, rather than collective competencies of teams or business units that they leverage in a truly agile manner to tackle day-to-day evolving challenges.
Tomasco (FUJIFILM): A specialized blend of IT, biotech and manufacturing learning initiatives are required to upskill existing teams and train our future workforce. The skills required for today’s manufacturing didn’t exist 5–10 years ago. We are working with our education partners right now to develop these cutting-edge training programs.
Of note, in North Carolina FUJIFILM Biotechnologies works with Wake Tech Community College to support the FUJIFILM Biotechnologies Early College Suite, offering a mix of IT and biotechnologies credits. In addition, we’ve partnered with WakeTech and other industry partners to create the nation's first-ever BioMechatronics apprenticeship. This specialized program offers development and training in mechanical and electrical engineering, motor controls, pneumatics, and programmable logic controllers (PLCs) used for industrial automation to prepare them to maintain and repair smart pharmaceutical and biotechnology manufacturing equipment and ensure it operates safely. This is how we prepare the future to lead our industry with Pharma 4.0 automation and innovation.
Aswath (Bionova): The transition to digital biomanufacturing requires a new skills mix, with a focus on key areas including data analytics and interpretation for process engineers and QA teams, automation systems training for operators, and AI/ML literacy programs for process development scientists
Bionova has begun cross-training staff in digital tools and process modeling, and partners with tech providers for specialized training in real-time analytics platforms. Cultivating this hybrid skill set is essential to keeping pace with innovation.
Seyf (Autolomous): With biomanufacturing growing more digital and automated, the industry faces a real need to adjust training of its people, not just to tweak existing workflows but to fundamentally rethink roles and skills. First and foremost, companies must map out the gap between today’s workforce and the digital future they want, especially in areas such as data analysis, automation, machine learning, and cybersecurity. Identifying who already has domain expertise and who needs new skills allows more disciplined training investments. This means building focused upskilling programs, certification tracks in automation or quality analytics, and hands‑on simulate‑to‑learn sessions on digital twins.
Partnerships with universities and technical schools can also help bring in new talent with the right mix of biotech and digital skills. Just as important is building confidence and engagement. People need time and support to adjust to new tools, and companies should provide practical, hands-on experience rather than relying only on theory. As automation and digital integration become standard, success will depend not just on new technology, but on how well people are prepared to use it.
Photo by Conny Schneider on Unsplash