The use of Artificial Intelligence (AI) is no doubt a hot topic in the biopharma industry. McKinsey has estimated the opportunity for generative AI in biopharma operations at $4 to $7 billion annually through workload and cost reductions, productivity gains, improvements to equipment effectiveness, and quality enhancements. We've also seen the use cases grow significantly across the product development lifecycle, including manufacturing automation.
We're often asked more specifically about use cases of AI in manufacturing, whether it's static or adaptive - where to start, how to deploy, and what's the real value. Or in other words, what's hype and what's reality?
These is one of the conversations of the ISPE 4.0 conference, where the theme of enabling effective digital transformation in operations rings quite true in today's biopharma landscape.
Use Cases and Benefits
While there is much buzz around AI usage, we must be careful not to see AI as a tool for all opportunities and problems. Rather, it's essential to properly pair the right challenge with the right solution and approach, whether that solution is AI or not. Our experience has shown that when the digitalization of operations is associated with an overall revision of workflows, it generates a significant improvement in performance, especially in terms of increased productivity (also due to the reduction of timelines) and compliance. Digitalization - in addition to allowing immediate availability of data, its transmission between the various functions, the constant verification of integrity, and the reduction of collection and transfer times - opens the door to advanced analytics from real-time performance evaluation to correlation statistics to the use of artificial intelligence for predictive analysis (because it is possible to reduce the failure rate and its associated costs and increase product quality). Benefits gained, include:- Predictive analyses (i.e., Maintenance, Laboratories, etc.)
- Product Quality Review by joining data from different sources (Labs, Quality Assurance, Production)
- Key performance indicators (KPIs) for continuous improvement (at Company, Site, Line level with OEE)
- Quality metrics (i.e., risk-based inspections from KPIs)
- On-demand analytics (i.e., for an audit, which user interacted with a given Electronic Record in a Quality Management System or Laboratory Information Management Systems)
- Simulation-based decisions (i.e., what to produce where, and scenarios including batch recall, etc.)
- Data reconciliation
Conclusion
There are a multitude of solutions available to clients on the market, but choosing the most suitable one is by no means a straightforward process.
Generally, a detailed analysis of the client's needs and the specifics of the context will help to identify the most cost-effective solutions and reduce the overall project costs and risks. Sometimes, changing the type or sequence of interventions or projects is enough to achieve savings without sacrificing results.
Authored by:
- Raffaella Vaiani
- Andrea Pieri