AI in Biologics CDMO Workflows: A BIO Conversation with Lonza

Why This Matters: Check out all our BIO 2026 coverage here.

At the 2026 BIO International Convention, Contract Pharma spoke with Zara Asgharpour, Head of External Innovation Integrated Biologics at Lonza. In the conversation at the show in San Diego, Asgharpour discussed the evolving role of artificial intelligence in biologics workflows.

She also discussed how AI is impacting the relationship between the technology industry and biotechnology players.

Lonza also released several other news items of note during the conference. The following is our exclusive Q&A with Asgharpour amid all of those developments.

Contract Pharma: What role does external innovation within Lonza play, and how does this fit into your broader biologics strategy?

Zara Asgharpour: The role of our team is basically to scout for external technologies. To look for what fits into the strategic needs of Lonza, and then set up collaborations. This is a full process managing the end-to-end, from scouting into making due diligence, into finalizing collaborations with external partners.

CP: How is AI transforming approaches for CDMOs, and what’s the impact of this on biopharmaceutical companies?

Asgharpour: I think AI is touch-basing toward the entire value chain, all the way from drug discovery into drug manufacturing. Also, into clinical trials, regulatory, and the post-clinical trial surveys. It’s an entire process, and I can speak to that it is basically touch-basing the manufacturing part. I think it is very important that we have harmonized data infrastructure already within the company to train the models. Because data is always the key.

Otherwise, modeling capabilities are something that our team certainly has. These are the approaches that we’re working on and constantly improving. Applying predictive modeling capabilities to being able to select the right process, the right clone, the right molecule, all the way up to commercial launch. It’s touch-basing the entire process, not just bits and bites of it. We already have capabilities within Lonza where we are applying the modeling in bits of the process. But a full chain needs to be to be there in order to get the greatest value out of AI.

CP: How will AI be the biggest value add for biopharma companies and CDMOs going forward?

Asgharpour: It is already in use in the drug discovery space. We see that it has brought a lot of advantages in cutting the timeline and making the selection process for your lead candidate a lot easier than it used to be. Also, it helps retarget the purpose of some drugs which maybe weren’t good for the purpose they were developed for. Then, finding new applications for them. This is already out there.

And then in terms of manufacturing, the greatest value that it will bring, of course, is the digitalization of our processes. Quickly telling us, and enabling the CDMOs, to select the right process. This is going to take the molecule all the way from the early stage into the commercial launch at the highest speed and lowest cost. I think this is the greatest value that AI will add.

CP: What are some of the opportunities and challenges in AI as it relates to drug manufacturing?

Asgharpour: AI is really the very top of the pyramid. If you look at the tech industry, it has taken them many, many years to gather all this data and build the models upon. In which way, they can then train the models meaningfully and use them for applications and purposes that they have.

Within the healthcare industry, whether pharmaceutical or medtech, medical devices, we are having a lot of challenges. The challenge in itself is not that you just get the AI and use the model, but we’re handling a lot of very sensitive information. Whether that is going to be about a new innovation or a new molecule, whether it’s going to be about a new process. Regulatory is involved, so this is a very strict and highly regulated environment. So these are some of the challenges and barriers to overcome.

And yet, you need to first lay the foundation before you can apply the AI, before you can even talk about the use of AI.The foundation should be there. The data should be there. The data needs to be standardized. Harmonized models need to be trained, and you need a vast amount of data in order to fit into the models. In order to get, basically, the right outcome out of it. So it’s a whole process. And AI is really the very, very, very top of it.

CP: Can you talk a little about the technical aspects of embedding AI into development and manufacturing workflows?

Asgharpour: We know how our existing processes work. We have gathered, over many, many years of working with our customers, a lot of data. They know our processes. And we also have a lot of data on how processes work within bioreactors, within manufacturing, etc. So how this will work is that you first need to, of course, have the foundation. As I said, the IT structure should be there, and so on. Then, you need to pick up the right elements that we want to predict, what makes sense, what doesn’t make sense. So if I have an existing process and I want to optimize it, I need to know what outcome parameters I am looking for, then feed my models based on those parameters.

I think embedding this within the process would involve these kind of steps that you need to fulfill before you can have a reliable outcome.

CP: Regarding AI’s utility as an experimental tool, can you describe the process of testing AI’s effectiveness under certain conditions?

Asgharpour: It’s a trial-and-error, and it’s a validation that needs to be taken in place. And like I said, it’s a matter of time, and it’s a matter of being patient. And it’s a matter of putting the right things in the model, because the model will just predict what you’re putting into it. So that means that you have to really be very careful about defining those outset parameters from the get-go.

CP: More generally, how can the partnership between the tech and biotech industries be enhanced?

Asgharpour: I think we are already seeing examples of it happening. A partnership between OpenAI and Novo Nordisk that was recently announced showed that tech knows a lot about how to treat the data. How to manage data, how to have the right infrastructure. They are the ones that are trying to set this entire process up in place. So there are great companies like Microsoft, obviously, and Google, among others. They have been working on all sorts of data for many years.

So I think that biotech and the pharmaceutical industry can greatly benefit from big partnerships in between. Because they do have the know-how, and they do have the knowledge on how to set up the basic parameters of getting there. And we do have a lot of know-how and knowledge on how the molecules work and how the pharma industry works, how the processes work. There’s a lot of learning between the two industries, and I think one cannot do without the other.