Skip to content

Interview

From Data to Knowledge

The Danish Environmental Portal (DMP) has set a clear direction for its work with AI, focusing on how the technology can turn environmental data into accessible, actionable knowledge for caseworkers and citizens. We spoke with DMP CEO Nils Høgsted about why the organization uses its AI strategy as a guiding framework—and how making data easier to understand and more relevant can help shorten case processing times and improve decision-making.

October 6, 2026

  • AI and Data

Nils Høgsted: AI Can Make Environmental Case Processing Faster and Better

“AI can be used in almost every solution we offer, which makes it essential to prioritize where it makes the most sense and creates the most value. We simply can’t pursue every opportunity at once,” says Nils Høgsted.

DMP is a public-sector partnership that brings together Denmark’s environmental data and makes it available to municipalities, regions, government agencies, and citizens. But the organization is very conscious that data alone does not create value.

“Our job isn’t simply to make data available. We need to make it useful, so it can be turned into knowledge and action,” Nils emphasizes.

That is why DMP has defined a clear strategic focus for its work with AI: making environmental data understandable, relevant, and practical for applicants and caseworkers.

From AI Strategy to Faster Case Processing

Asked where AI can make the biggest difference today, Nils points to two specific areas: environmental assessments and environmental permits.

Cases in these areas are often lengthy and complex. They involve multiple public authorities and require numerous criteria, guidelines, and regulations to be considered. Gathering and clarifying all the necessary information can take years.

“Every iteration between an applicant and a public authority takes time. Complex applications may go back and forth several times before the information is considered sufficient for the case to be processed,” Nils explains.

Lengthy processing times delay investments and create uncertainty for both businesses and citizens. This is where Nils sees a very tangible opportunity for AI.

“If AI can make it easier to identify what information is missing and which data is relevant to a specific case, we can reduce the number of iterations and significantly shorten processing times. That strengthens Denmark’s competitiveness,” he says.

First Step: AI-Assisted Search

The first—and most mature—step toward AI-assisted case processing is making data easier to find and use. For Nils, the immediate opportunity is clear.

“Right now, we see the greatest impact from AI-assisted search, which allows us to bring data trapped in static PDFs to life,” he says.

In practical terms, “bringing data to life” means users can ask questions in everyday language and receive relevant excerpts that point directly to the information in the document.

DMP has already integrated AI-assisted search into several of its solutions.

One example is EA-Hub, a shared platform containing thousands of environmental assessments and decisions. AI-powered search makes this information accessible in a new way, enabling caseworkers and advisors to find exactly the knowledge they need faster.

Another example is DMP’s Danmarks Arealinformation and Arealdata platforms, where synthetic metadata and AI-powered search help users find relevant information on issues such as soil contamination, building restrictions, and protected areas.

“With AI-powered search, you can ask which data is relevant to a particular construction project and get a much more focused dataset to work with,” Nils explains. “You also don’t need to know that an aerial photograph is called an ‘orthophoto’ in technical terminology. You can search using everyday language, even with typos.”

From Search to Generative AI —Applying Rules and Criteria

“Once it becomes easier to find the right data, the next step is moving from search to generative AI, where we start connecting data with rules and criteria,” says Nils.

He uses building permits to illustrate the principle.

“If you want to build a house in an allotment garden, one criterion might be that the maximum building coverage is 15%. AI can search the data to determine whether that criterion has been met. Once you can check the criteria one by one, it quickly becomes clear whether a case can be approved, or what information is still missing before the case is sufficiently documented. Over time, that could provide the basis for draft decisions.”

“The same logic can be applied to environmental assessments. One criterion might be whether a protected species, such as the moor frog, has been recorded in the area. If it has, the next step is to identify which mitigation measures could be used.”

DMP is developing an AI-based tool that can work across different types of cases, bringing rules and requirements together in one place and connecting them with the data needed to document compliance.

But with every increase in complexity comes a greater need for data quality, controls, and human validation.

As Nils puts it: “AI-assisted search creates value without introducing the same level of risk, while generative solutions that produce draft decisions require much tighter governance and continuous quality assurance.”

Turning AI Strategy Into a Management Tool

Today, DMP’s AI strategy serves as a shared compass that makes it easier to set priorities and maintain focus. But the strategy grew out of practical experience, Nils explains.

“We started by experimenting and only built the structure once we had enough experience to scale responsibly.”

The strategy is built around clear principles, a structured decision-making process, and a dedicated domain group that ensures AI initiatives support business objectives rather than simply emerging because someone has a good idea.

One practical measure is the direct connection between strategy and backlog, making the strategic direction visible in everyday work rather than something that is only revisited at the beginning of the year.

For Nils, success ultimately comes down to measurable savings.

“At the end of the day, it comes down to measurable financial impact,” he says. “AI-assisted search is significantly less expensive than traditional search. At the same time, we’re seeing annual savings of around 10–20% because our vendors use AI for maintenance and development work.”

Because the results are measurable, DMP also follows a practical principle: solutions must be able to scale without costs getting out of control. A solution is put into production, usage is monitored, and capacity is increased from there.

“We avoid building generic solutions for everyone. Instead, we focus on the areas where the potential value is greatest and where any additional costs can be covered,” Nils explains.

Nils Høgsted

The Best Solutions Are Built Together

Another important principle for Nils is that no organization can — or should — try to solve everything alone.

AI is evolving rapidly, and the uncertainty is real.

“We genuinely don’t know whether what we’re developing today will still be the right solution four months from now,” Nils says.

That is why DMP actively pursues partnerships. Not only to share costs, but also to validate solutions and strengthen its position in a field that is constantly evolving.

“We’ve started working with the Department of Computer Science at the University of Copenhagen and KOMBIT to develop shared software components that can be used across organizations. There’s no reason for each of us to reinvent the wheel,” Nils says.

Reusing both experience and technology components allows the entire sector to move faster.

The Technology Is Ready. Regulation Needs to Catch Up

Asked what he hopes AI-assisted environmental administration will look like five years from now, Nils has a clear vision: shorter case processing times, more consistent decision-making, and an environmental administration where more issues are clarified earlier in the process.

Among other things, he hopes developers will be able to consider environmental requirements during the planning phase, avoiding situations where a project is planned for an unsuitable site in the first place, and enabling permits to be issued much faster.

“The potential of AI is enormous,” Nils says. But he is equally clear about what ultimately determines the pace of progress.

“The biggest barrier isn’t the technology. It’s recognizing just how significant the transformation brought by AI really is and being willing to embrace it with our eyes open.”

He emphasizes that clearer rules and a more unambiguous regulatory framework are essential if automation and more efficient public administration are to keep pace with technological development. When legislation has not been adapted to a digital reality, it naturally limits what can be automated, regardless of how capable the technology becomes.

If Denmark is to realize the full benefits of AI across public administration, the willingness to transform must also be reflected in the frameworks within which public authorities operate.