Thomas Halfmann, smart factory advisor at eschbach, and Andreas Eschbach, CEO and founder of Eschbach, outline five steps manufacturers can take to break down these barriers and realise the vision of the adaptive plant.
Shutterstock - Cagkan Sayin
While digitalisation efforts are ramping up across the face of pharmaceutical manufacturing, most companies are still being held back by information silos and disconnected systems, which are preventing them from achieving a fully digitally mature, adaptive plant.
If an entirely paper-based plant represents one end of the spectrum for digital maturity, then the adaptive plant is the other – a facility with full value chain integration, powered by AI-enabled insights, data-driven automation, and real-time process verification, control, and adaptation. However, the stark truth is that most pharmaceutical manufacturers are actually closer to the former rather than the latter.
Mapping the digitalisation journey
Many in the industry will be familiar with the Digital Plant Maturity Model. Developed by BioPhorum Operations Group in collaboration with leading biopharma manufacturers, the model serves as a barometer for digital maturity, and breaks down into five levels. From least mature to most, these are:
- Pre-digital plant
- Digital silos
- Connected plant
- Predictive plant
- Adaptive plant
At present, most pharma manufacturers sit somewhere between level two and three. At level two, individual systems – such as electronic batch records, maintenance logs, or shift reports – may be digitised, but these systems are disconnected and inaccessible, hence ‘digital silos’.
At level three, there begins to be some integration across operational systems, such as the ERP and MES. Digital workflows begin to be established, and data collection becomes a key capability. However, this still has its limitations. System interfaces are not standardised and the effort for implementing and maintaining a consistent data architecture is considerable high. Data flows between systems but often lacks contextualisation. While the plant is now connected, it is far from predictive or adaptive.
Some manufacturers are beginning to move into level four, where integrated systems and advanced analytics provide better visibility across the organisation. These plants are starting to use real-time dashboards, predictive maintenance tools and exception-based quality review to anticipate issues and make smarter decisions faster. However, even here, systems support decision-making, but they don’t yet adapt on their own.
A truly adaptive, level five pharma manufacturing facility is a different entity altogether. It goes beyond connectivity and prediction to create an environment that is intelligent, responsive and self-improving.
This would be typified by autonomous, real-time process optimisation using AI and predictive analytics, coupled with seamless data integration across manufacturing, quality, supply chain and enterprise systems – all accessible at the right levels.
In addition, the adaptive plant has modular, plug-and-play infrastructure that enables flexible scaling and rapid adaptation. Other characteristics include maintenance and quality systems that are self-directed with minimal human intervention, with the frequent use of digital twins and simulation tools to support continuous learning and decision-making.
Information is power
The industry isn’t there yet, but the building blocks are already taking shape. The gap between today’s reality and tomorrow’s adaptive plant isn’t just about technology; it’s about how information flows.
Software like eschbach’s intelligent operations platform Seqonis (formerly known as Shiftconnector) is the operational backbone of adaptive manufacturing. It centralises critical operations like shift handovers, event logging, task management, and KPI tracking into one seamless platform, eliminating information silos and harmonising data to providing real-time visibility across the plant.
Crucially, it also has the ability to transform data into action. In an adaptive plant, data must be more than stored — it must be accessible, searchable, and instantly usable to empower teams with real-time insights, cross-shift visibility, and faster decision-making. This begs the question – how can it be used to realise the Level 5 adaptive plant?
Five key steps to develop your own adaptive plant:
1. Assess your digital maturity
The foundational step in moving towards the adaptive plant is working out where you currently stand. With the engagement of all major stakeholders, conduct an as-is assessment and create a heatmap of how your plant tracks against the criteria outlined in the DPMM.
2. Define your strategy, architecture, and roadmap
Now you have a clearer idea of where you are, you can use this knowledge to construct a roadmap of where to go next. The key here is aligning your digital goals with your business needs – approach this challenge pragmatically and balance your ambitions against financial reality. Target the low-hanging fruit which will deliver value quickly, rather than aiming for 100% digitalisation.
3. Implement an intelligent operations platform
Introducing an intelligent operations platform is the key to advancing beyond level three of the DPMM, which is where many manufacturers reach a ceiling. Systems such as Seqonis not only ensure that your data is connected, but harmonised, consolidating it into a single source of truth that can offer 360º real-time visibility of critical information via smart dashboards.
At Bayer AG Drug’s Substance Finishing Plant, for example, over 20 disparate communication tools were consolidated into one single source of truth with the intelligent operations platform Seqonis, fostering greater clarity, collaboration and data integrity across teams and shifts.
4. Empower your workforce
As part of your digital transformation, operators on the shop floor should become decisionmakers, not task executors. A truly adaptive plant is one that leverages the implicit knowledge of its operators and converts it into explicit systems, introducing AI tools to support – not replace – human expertise.
For instance, our Seqonis Artificial Manufacturing Intelligence (SAMI) Chat function acts as a conversational AI akin to ChatGPT, but based entirely upon your plant’s data. For instance, users can ask ‘How much delay did the laboratory cause in the past?’ and SAMI Chat will respond as if talking to a co-worker, generating a summary and citing the relevant shift logs and incident reports. With the ability to quickly sift through plant data in this manner, operators on the shop floor are able to make more informed decisions.
5. Scale across boundaries
Once your plant is fully adaptive, this same strategy can be rolled out across all of your domains and sites. Digital practices can be standardised across multiple manufacturing sites or CDMO partners, while enabling data sharing can help foster even greater process transparency across departments, sites, and companies.