Swiss-headquartered contract development and manufacturing organization (CDMO) Lonza is making progress in leveraging artificial intelligence (AI) and advanced analytics as the company looks to unlock the power of more than 40 years of process data.

In a Monday keynote address to the annual meeting of the International Society for Pharmaceutical Engineering, Lonza CEO Wolfgang Wienand said AI will “play a pivotal role” going forward in how the company manufactures pharmaceuticals and ensures quality, as well as creates supply chain efficiencies.  

“Lonza has a great starting point because no one else in the pharmaceutical industry has brought more drugs to market — more than 140,” according to Wienand. “We have been working on 3,300 clinical molecules, 1,500 tech transfers, and 32,000 batch records.”

Wienand said the company is “cleaning up” this vast amount of data to make it usable and through the “Lonza intelligence layer” is leveraging AI to gain actionable insights that improve drug development, manufacturing efficiency, and decision‑making. At the same time, he emphasized that the CDMO’s implementation of AI is “always with a human in the loop” and the company applies its data responsibly and with integrity.

At Lonza’s Capital Market Day last week, Wienand said the company has an extensive library of pharmaceutical manufacturing data, building on more than 40 years of experience. The goal is to tap AI and advanced analytics to build a “cognitive CDMO” connecting scientific knowledge with operational data to learn, predict, and recommend.

Weinand’s presentation pointed to a newly established Lonza AI Factory, responsible for building, scaling, and governing AI across the company and integrating analytics, machine learning, generative AI, and agentic AI with Lonza’s decades of domain expertise.

“We are the one company in the world, including large pharmaceutical companies, with the highest number of executed batches,” Wienand said. “We are sitting on the richest database in terms of manufacturing data of pharmaceutical drugs.”

Gordon Bates, Lonza’s head of integrated biologics, told analysts that the CDMO is assessing where AI can optimize workflows and provide benefits from a workforce utilization perspective. The objective, he said, is “scaling this company to make sure that we can use our great people who focus on what they’re brilliant at” while “not tying them up in things that are transactional, where potentially we can use AI to help do that.”

3 core AI capabilities

Lonza is building towards three core capabilities: Al-accelerated development and tech transfer to reuse knowledge across molecules to select experiments and anticipate risk; Al-guided manufacturing to predict performance and recommend targeted interventions to reduce variability and increase yield; and AI-enabled quality to detect signals earlier and strengthen proactive quality management.

According to Wienand, Lonza’s AI-guided process optimization has already successfully turned complex batch histories — more than 100 historical batches across reactor conditions, raw materials, feeding profiles and production outcomes — into targeted actions for a large pharma customer by identifying process-signal combinations that drive titer, yield, and process performance.

“Together with the customer, based on a number of historic data, we have been able through AI to actually increase the output by 15% by identifying relevant signals in historical batches between process parameters and output through a human-in-the-loop, [Manufacturing Science and Technology] expert, make decisions which led to significantly improved output,” Wienand said.

William Blair analysts in a note to investors last week described Lonza as being in the “early stages of building an intelligence layer that will leverage its more than 40 years of process data to power its vision for a ‘cognitive CDMO,’ using AI to accelerate development and tech transfer, optimize manufacturing, and strengthen quality management.”

For now, the analysts said the early results are “promising” given the customer example of AI-guided process optimization achieving a 15% increase in output per batch, which highlights the “potential to improve yields, consistency, and manufacturing efficiency over time.”