Q1. Could you start by giving us a brief overview of your professional background, particularly focusing on your expertise in the industry?

I began my career at Danske Bank, focusing within the field of analytics for cybersecurity analysis. I then moved into the renewable energy sector, focusing on digital strategy and data governance / data management initiatives. 

Now, as Data Governance Lead in Enterprise IT, at Novo Nordisk, I create, implement, advise and guide standards and frameworks across business units, optimizing processes, SOPs, and oversight for robust, compliant data practices, company-wide. These standards mean creating clear roles, responsibilities, and procedures that secure data quality, consistency, and control, while also adapting to the needs of individual business entities. My role is both global and flexible — I guide change management, ensure cross-unit collaboration, and drive the enterprise data strategy with measurable impact, maintaining tight connections with leadership and operations for effective implementation and value realization.

 

Q2. How do you see AI transforming traditional data governance frameworks in large enterprises over the next 3 to 5 years?

There are two main components to consider in this question. First, it is important to define data governance clearly. 

Data governance involves ensuring that the right people have the appropriate roles and responsibilities to manage the business processes that support the organization. The aim is to have the right people running the right processes, which helps the business work more efficiently and improves data quality. AI can serve as a valuable tool here, supporting both people and processes. In this way, AI can make governance activities more effective.

AI will identify potential areas for process improvement

Currently, setting up data governance typically involves direct discussions with business users to understand their processes and identify areas for improvement. These conversations often reveal bottlenecks and ways to improve. Because documentation is sometimes lacking, we often create or update it during these sessions. AI can make these workshops more efficient. With good documentation, AI can quickly identify areas for improvement, saving time and reducing the need for lengthy manual workshops.

Maintaining and updating the data used by AI systems

It is important to note that AI itself relies on data, which must also be governed to address issues such as potential bias. If outdated or incorrect business processes are used to train corporate AI systems, organizations risk becoming increasingly reliant on inaccurate outputs. There is a risk that users may accept AI-generated recommendations at face value without verifying their accuracy, especially if the underlying business processes are outdated.

This highlights the importance of data governance in assigning responsibility for maintaining and updating the data used by AI systems. Key considerations include identifying who is responsible for updating AI data, determining the frequency of updates, and understanding the risks and impacts if this is not done. Data governance must be applied to the data that powers AI. This governance should be in place both before and during the use of AI. At the same time, AI has the potential to transform data governance by accelerating interviews, analysis, and information retrieval.

 

Q3. What emerging technologies or innovations beyond AI should we watch for that could further disrupt or enhance enterprise data management?

I believe the most significant change in AI currently extends beyond the typical tools and everyday applications, such as ChatGPT or Microsoft Copilot. As of today, a lot of focus goes into generative models that can generate text, image, or even video information for you. Those are nice use cases, but they are reactive use cases – since the AI requires an input from a user firstly. It provides an answer only when you ask it a question. That's a reactive approach. But I think the future is moving in two new directions.

Proactive AI

Proactive AI can learn how you work, your schedule, your business processes, and the people you interact with. It can then ask questions for you or, for example, send you daily updates with the latest news from the business areas you support. It could also help you prepare the documents you need for your meetings each morning. This is just one example of how proactive AI could help, even beyond enterprise data management.

Agentic AI

Agentic AI refers to a collaborative system where different AI systems work together, each highly specialized to solve specific problems. This could change not just enterprise data management, but how companies operate overall. These AI agents will require high-quality data to perform their jobs effectively, which is where data governance becomes crucial. They could also help connect data across departments, making it easier to use information that is currently stuck in silos. Commercial data can be understood from various perspectives, including finance, supply chain, commercial excellence, and sales management. Each of these business roles and personas has its own individual needs.

This is how I see AI moving forward. It's not just about new technology but about using AI in ways we haven't seen before. In the next few years, I think we'll see big vendors with reliable products start using AI in areas that have always required a lot of manual work, like master data management. Currently, these tools assist data stewards in managing and improving data quality, identifying duplicates, and detecting errors. 

Most of the current tools use simple logic or rules set up by people. With AI, these systems can automatically spot patterns, find duplicates, or catch out-of-range data, eliminating the need for a data engineer to set up each rule. AI could even alert data stewards to new quality issues as soon as they appear.

Besides that, in each of these tools, whether it's a Data Catalog or an Information Modelling tool, applying AI inside it will be significant, due to the scale at which these vendors' systems are being used.

We can already see this with tools like Microsoft Copilot. It's a great tool, but at scale, it still has not made the full impact as an integrated agent across all your files and Microsoft products. If vendors can make this work well at scale, it will be a major innovation and a real game-changer for businesses.

 

Q4. How do you assess the competitive positioning of traditional enterprise software giants versus pure-play AI governance startups in this market?

There are many pure AI governance start-ups, but their current activity doesn't guarantee they'll still be around next year. This is true for all start-ups, but especially in AI. There's a boom right now for finding fast cash, and it's a bit of a gold rush for people to use AI, partly because some people have started using AI itself for startup creation and management. Some people think they can work less and still be part of a fast-growing field. However, you need to be cautious when working with these companies, as some may take shortcuts, particularly if they lack a strong team of developers or cybersecurity experts. Ensure the infrastructure is ready to protect against threats such as DDoS attacks, various hacking methods, and zero-day vulnerabilities.

Cyber threats like these need to be taken very seriously. That's why companies like Novo Nordisk, or any major player, have a strict due diligence process before working with a start-up. Let's say that an AI governance company is well established, and it passes the due diligence process. In that case, a strong business case will be discussed on whether to integrate this AI governance tool. We might choose to add the start-up's tool to our systems, or we might go with a major vendor, which usually costs more but can guarantee strong service-level agreements, reliability, and long-term support. The company's establishment within the market and its pure expertise is sometimes more valuable than fast innovation, although ideally people want both. AI governance startups can be competitive if they do things right, and think thoroughly through all aspects, as well as hire experts in the industry they are competing in, instead of relying only on AI. 

AI governance is a tough area, with only few laws or regulations so far. The EU act has been passed, and came into force in a delayed activation manner, where some provisions are already in application, and some will activate during 2026. In the US, there is no comprehensive federal AI statute, however there are binding measures across a couple of states (NY, Colorado), with more to come. Japan is also starting, with the AI Promotion Act. However, generally, AI governance remains an unclear area for unspecialized actors. So, with a startup, you will need to be very clear on your business proposition and how you can help, since there are very few people who have done it. 

To sum up, I do believe there is potential here. If executed well, these start-ups can provide rapid support to companies in need, even if progress is sometimes slow. However, gaining real expertise in this area is difficult, so companies need to be very thorough when choosing a partner.

 

Q5. What are the current gaps in the market for AI and data governance solutions that represent significant growth potential?

There are major gaps in the current market for AI and data governance solutions, with the most significant being the disconnect between what traditional vendor solutions offer and the true capabilities of AI. For example, a typical Data Catalog tool has offered some of the same core functionality for years. 
Established vendors usually introduce AI features cautiously, with limited initial capabilities. It's rare for them to launch AI that automates or overrides most user tasks. Their caution is justified: overly powerful AI could accidentally modify or delete critical data—posing serious risks if mistakes occur. As a result, most vendors currently limit AI's access and scope to avoid potential damage.

Today's AI is far more advanced than older NLP tools or basic chatbots, yet its use is still tightly restricted. The key market gap is creating ways to unleash AI's productivity, safely. This means providing controlled environments, like sandboxes, where AI can operate, make changes, and drive efficiency, while always allowing a rollback to the original state if needed. Features like version control, sandboxing, and cloud-based backups are becoming essential, since they enable organizations to benefit from AI’s capabilities without risking data loss or unintended changes.

 

Q6. If you were an investor looking at companies within the space, what critical question would you pose to their senior management?

If I were considering a partnership in this space, my first question would be about the depth of in-house human expertise relevant to the problem they are tackling—especially legal, compliance, or domain-specific knowledge for AI governance. 

Next, I would ask about their AI training datasets & methodology: 

• What type of data was used? From where was it collected?
• Is there comprehensive documentation or a data sheet? 
• How do they assess and address bias in their data sets? 
• What criteria guided their training process? What data was left out - and why?

I would also question senior management about how they defined the AI’s constraints and freedoms. Finally, I’d ask for references, partnerships, case studies, or a portfolio that demonstrates real-world impact. Maybe this is the first company they're collaborating with, and the start-up might not have references. In that case, I would clearly ask them: “Do you have clear case studies, case examples, or any portfolio that you are working with, and then have you been able to demonstrate value from?”

I would ideally want to see both qualitative and quantitative evidence of business value. The key metrics I’d look for are time saved, measurable improvements in data quality, and cost redistribution (not just cost savings). In practice, governance rarely increases the budget pool per-se, but it enables teams to reallocate time to higher-value activities—for example, optimizing a two-week (10-days) process into a 4-day one, thereby freeing up staff for more strategic work.