The Biggest Gap: Technological Enthusiasm Without Problem Focus

In our interview, Dr. Herwig describes a fundamental imbalance. For many years, the spotlight was on technology rather than on the measurement task itself.

In the early days of PAT, there was a widespread belief that purchasing a spectrometer was sufficient to implement PAT. The result was significant investment, unmet expectations and, ultimately, a rather disillusioned market. Even today, professional discussions often revolve around measurement principles and technical possibilities instead of industrial reality.

Herwig summarizes it clearly: “We talk too much about instrumentation and too little about the concrete pain points PAT is supposed to solve.” The gap between theory and practice emerges when technology is not embedded in a coherent strategic framework. Without a solid understanding of the process, PAT remains an isolated tool.

Why SMEs Hesitate and Why They Don’t Need To

While large corporations operate dedicated PAT departments and benefit from economies of scale, small and medium-sized enterprises (SMEs) often shy away from the perceived effort. However, Herwig sees a misunderstanding here. PAT does not necessarily have to be expensive.

He refers to the broader definition in ICH Q8. PAT includes not only hardware sensors, but also software solutions, soft sensors and digital twins.

Many SMEs still associate PAT with costly spectroscopic systems and overlook the fact that even simple, appropriately timed measurements can provide crucial information. In slower processes, such as cell culture, measurements every four, eight or even 24 hours may be entirely sufficient.

Herwig’s recommendations are clear:

1. Clarify the definition. PAT is more than inline process analytics.
2. Prioritize the measurement task. What information is truly required to make sound decisions?
3. Leverage soft sensors and digital models. They are cost effective and often highly impactful.

The Next Five Years: Prediction, Control and New Product Worlds

The pharmaceutical landscape is changing rapidly. New modalities, ATMPs, individualized therapies and multiproduct facilities are shaping the future. With them comes a growing need for predictive capability and robustness.

Herwig emphasizes that variability, especially in biological starting materials, cannot be eliminated. It must be measured, understood and integrated into process control.

As a result, PAT will

• be used more strongly for predictive control.
• be closely linked with digital twins.
• become indispensable for continuous manufacturing.
• benefit significantly from artificial intelligence, particularly in transfer learning and the improvement of calibration models.

He highlights the importance of end-to-end thinking. Process models that link data from starting material to final product will become key tools for robust manufacturing.

PAT for ATMPs and Biopharma: Variability as the New Normal

Unlike conventional pharmaceuticals, ATMPs such as cell and gene therapies involve highly variable raw materials. Each patient brings different starting conditions.

This is where PAT demonstrates its greatest strength. It enables the measurement of initial attributes and allows processes to be adjusted for individual patients, often with only a single production run available.

In this field, predictive and data-driven control is not just beneficial. It is essential.

The Role of Conferences Such as PAT Connect 2026

Professional events are crucial for progress, provided they remain close to practice. Herwig advocates formats that

• are driven by real-world industrial challenges.
• actively involve regulators to build understanding and trust.
• offer hands-on workshops that work through real use cases.
• make business cases visible instead of presenting polished slide decks only.

Only through collaboration between industry, technology providers and authorities can PAT reach its full potential.

Sensor Selection and Process Understanding: The Right Order

One key message runs throughout the interview: “Do not choose the sensor first. Understand the process first.”

Sensor selection follows from:

1. process understanding
2. observability analysis, meaning which variables the model truly requires
3. signal to noise ratio and measurement frequency

He strongly cautions against generating measurement points instead of information points out of technological enthusiasm. Measurements only create real value when they contribute meaningfully to understanding.

Conclusion: Rethinking PAT – Pragmatic, Process Driven and Future Ready
Dr. Christoph Herwig makes it clear that PAT is far more than technology. It is a strategic approach to using data for better decision making.

For PAT to move from pilot project to lived reality, it requires

• demystifying the term,
• refocusing on the measurement task,
• stronger integration of digital methods,
• predictive and robust control strategies,
• and platforms such as PAT Connect that bring together knowledge, practice and regulation.

The transformation has already begun. PAT stands at its center. The coming years will show which companies seize this opportunity and make proactive process design the new normal.