Driving Data Innovation and AI Adoption in Clinical Trials: A Q&A with Novo Nordisk’s VP of Data Systems Innovation

With two decades of experience spanning data management and clinical trials, Ibrahim Kamstrup-Akkaoui, vice president of data systems innovation, Novo Nordisk, has built a career at the intersection of science and technology. Now leading several of Novo Nordisk’s major clinical systems, he is focused on driving innovation, improving data handling, and accelerating trial processes to bring medicines to patients faster—all while maintaining the highest standards of quality and compliance.

In this Q&A with Applied Clinical Trials at the 2025 Veeva R&D and Quality Summit, Kamstrup-Akkaoui discussed his role, the company’s DataNow program to simplify system use for trial sites, and how Novo Nordisk is taking a measured approach to leveraging artificial intelligence (AI) and automation across the clinical development lifecycle.

ACT: You have 20 years of experience in data management and in clinical trials. Can you tell us more about what your current role at Novo Nordisk entails and what you're focused on?

Kamstrup-Akkaoui: Currently, I'm responsible for several of our major clinical systems, and with that, of course, comes a lot of development and innovation. All in all, you can say it's for supporting how we run our clinical trials, how we collect our data, how we handle our data, and trying to speed up processes, because we are constantly aiming at getting products to the market faster and being available for our patients. That is the ultimate goal. Of course, there are several layers in that. There's also some quality aspects and compliance aspects that we need to consider. I think this is where it means a lot that you've tried a few things within this industry.

I started my career in pharma by working on a lot of clinical trials, starting off with Phase I trials, where you get a lot of repetition because they're shorter, usually. The process is pretty much the same and provided a lot of good experience that I've been able to transfer to bigger trials. Coming from an IT (information technology) background gave me a little bit of an advantage on how you can apply technology to do clinical trials from a system and data handling point of view. Of course, having accumulated experience over the years, I've gotten more responsibility within those areas and got me to where I am today combining both worlds, my passion for science and experience with technology.

ACT: You have a program at Novo Nordisk called the DataNow program. Could you share more info on the objectives and the goals of that program with us?

Kamstrup-Akkaoui: The fact is, we give the clinical sites a lot of systems to work with, and I hear numbers between 15 to 20 different systems that we as Novo Nordisk give the sites to work with on a clinical trial. Considering that some of these sites work with other sponsor companies in the industry, then you can probably multiply that number by two or three. At the end of the day, that gives us a lot of challenges, because it's a very complex environment to navigate and use, so mistakes could happen if you don't put enough effort into your training and education for the staff involved in the clinical trials. That's both from the sponsor side and the clinical investigator side.

Knowing that, and receiving that feedback over the years directly, because I was representing that part of the business on the clinical trials, I thought if I now get a responsibility and a mandate to make this part of the world a better place, what can we do?

This is where the DataNow journey began. We wanted to take a user-centric approach all the way from the sites, the CRAs (clinical research associates), and other good functions in the clinical development process, and take a look at how we can simplify the system landscape. We are working towards giving the users a platform where the patient journey throughout the trial is built into a process on top of that platform.

ACT: How is Novo Nordisk leveraging AI and automation to streamline clinical trial operations, and which parts of the trial lifecycle are showing the greatest impact so far?

Kamstrup-Akkaoui: I would say we're still early days on the AI piece. When we started wanting to apply AI to different activities in our clinical development process, we realized that starting small was a good option, and then figuring out. It didn't take a lot to do RPA (robotic process automation) solutions, rather than, if you want to do real AI, which would take a different skill set and would probably take a little bit longer to get the solution in place. Also, the validation part of the implementation or development was an unknown factor a couple of years ago.

For now, we are starting with the small solutions. We're doing things with AI around testing and validation of our systems. Now, what we're working with Veeva on at this time is also using AI to actually set up our systems. Then, for me, the next steps would naturally be trying to handle the data that we collect in clinical trials and replacing the manual effort there with AI algorithms. I think this should very much be possible, but we need to get the experience working with AI while not influencing the impact and outcome of a clinical trial or clinical program. I think it's a wise approach we're taking now, where we are applying the technology on pieces of the process that we know really well and can control or stop in due time if things are not going as expected. Looking ahead, I think applying AI to the data handling part is going to be really interesting.

ACT: What best practices would you recommend for clinical operations teams looking to introduce AI-driven tools without disrupting existing workflows or trial timelines?

Kamstrup-Akkaoui: I still see a lot of manual activities and processes in the clinical operations space. Let's take an example that could be similar to the data handling challenge, there's still a lot of manual handling of protocol deviations. As one example, this is a place where there is a box of activities to be handled with AI algorithms, because you really want to have them to focus on running the trial well. If they get the too tied up with all this manual effort of dealing with the protocol deviations, that’s going to be a challenge.

Using AI to identify or help identify where the focus of the personnel should be, rather than trying to look into all elements of a trial, like revenue or data. Then having AI handle what it can handle, and what it cannot handle it identifies for you. We can do part of it today with some basic technology related to reporting and automation, but I think AI can take it to the next level.