At the Copenhagen headquarters of renowned architectural practice Henning Larsen, the launch of the firm’s new enterprise partnership with Krea began with a clear premise: AI should not be reserved for a small group of specialists. It should be tested, questioned and shaped collectively.
Henning Larsen have chosen a distinctively bottom-up approach, letting architects across the firm explore and use any AI tools before committing to an enterprise partnership. The goal is to ensure that new technology adoption doesn’t get confined to tech-savvy users, a classic trap many enterprises fall into: a group of power users absorbs the tool, and the rest of the organization simply delegates tasks to them in the spirit of “automation.”
For Eliana Nigro, Head of Digital Adoption at Henning Larsen, the partnership with Krea reflects a broader culture of shared knowledge. She opened the event by describing the firm’s approach to AI adoption in the “spirit of collaboration” and as a form of “collective learning.” Rather than concentrating AI expertise within a single, siloed team, Henning Larsen is pursuing a horizontal model, where designers across the practice, from rendering to video to urban planning, help define how the technology should be used.
A visualization of a planting pallet created with Krea by landscape architect Brian Mallig Collado. Henning Larsen, 2025
This approach brought an unexpected theme to the surface — not the story of AI automation you might expect, but something far simpler and more essential: communication.
Henning Larsen’s belief is that when every member of a team gains access to a tool that enables efficient visual communication, something shifts. Now, they are not just working faster; they begin to understand each other better.
AI as a Tool for Communication
Across presentations by Henning Larsen architects Joachim Makholm, Samuele Agrimi and Martin Petersen, AI was shown less as a machine for producing finished proposals and more as a way to make ideas visible earlier in the lifecycle of a project.
Makholm described AI as an “additional production tool embedded in the design process” and “a supportive tool to help us convey our ideas better.” In one case, where the team could not use a drone to capture site imagery, Krea was used to generate visuals that helped guide early client conversations about the place.
Krea is used in the office to explore methods for enhancing low-resolution aerial imagery. Pictured here: a non-project test using Google Earth imagery from Vík, Iceland. Nitsan Bartov, Henning Larsen, 2026
Petersen focused on AI’s ability to align perception across teams and clients. Using a single base image, his team generated distinct programmatic scenarios that depicted different interpretations of a space as a starting point for discussion. Work that once required hours in Photoshop could be produced quickly enough to inform the conversation in real time.
Architect Martin Frank Petersen used Krea to create early-stage visualizations of a multi-functional space in different scenarios. Pictured here in animated form: The empty base image, a visualization of the space used for a clay workshop, and finally, a visualization of the space being used as a community dining hall. Henning Larsen, 2025
For architects, that speed fundamentally changes the role of the image. Instead of arriving at the end of a design phase as a polished rendering, imagery can become a working instrument inside the process. It can clarify intent, expose disagreement and help clients understand what a particular conceptual direction might feel like without it needing to be fully modeled.
Nigro described this as a shift from presentation towards co-creation. “It’s no longer about persuading someone, but finding ways to agree and move in a certain direction,” she said.
The shift towards co-creation is a pattern Krea’s Partnerships Team sees across the industry. “The firms getting the most from these tools aren’t using visuals as a single sign-off moment; they’re using them to open up a conversation,” says Helena Grau. “When a concept can be understood from the start and shaped in real time, you stop trying to align on a vision and start building one together.”
Landscape Architect Philippe Larocque used Krea to create visualizations of different flooding scenarios. Pictured here, the base scenario (top left), a visualization of a new creek (top right), then a visualization of a small flood (lower left) and finally a scenario visualization of a large flood (lower right).
In landscape and urban design, this approach has the potential to be especially powerful, allowing teams to test designs against environmental conditions, atmospheric qualities and lived experience much earlier in the process.
Crucially, that power is conditional: the value comes only when these tools are in the hands of people who understand how climate actually behaves — domain experts who can read the outputs critically and validate them against reality, not AI specialists taking results at face value.
Why Krea Matters
Diego Rodriguez, Krea Co-Founder and CTO, positioned Krea as more than another AI image tool. He described the company as an attempt to build something closer to a new Autodesk or Adobe: a creative suite for an AI-native era.
He opened with a candid line: “People are drunk in the AI industry; hopefully I can bring a bit of sobriety.” AI is often discussed through extremes within creative industries, from over-the-top hype to aggressive dismissal. As an antidote to this, Krea’s aim is to build tools that make practical sense for design professionals.
Krea conducts its own foundational model research, which Rodriguez presented as a key distinction from products that simply wrap existing models.
Krea’s Co-Founder and CTO Diego Rodriguez demos Krea at Henning Larsen’s Copenhagen Headquarters. Photo by Isuru Perera
The event also coincided with the launch of K2, Krea’s new model that allows users to train their own models and collaborate with them. K2 has recently been top-ranked for text-to-image models by Artificial Analysis, a landmark achievement for an insurgent AI lab breaking into a ranking otherwise dominated by OpenAI and Google.
For architecture firms, the ability to develop models around their own visual language, professional experience and workflows could become increasingly important, and the ability to build domain-specific tools around those models is where it gets genuinely interesting for the industry.
Working with Henning Larsen, Rodriguez said, “makes us better at what we do: building the most powerful generative tools possible and ensuring Krea tooling stays above the high bar that architects are known for.”
The Risk of “Convincing” Images
One of the most compelling moments of the event was when the speakers addressed some of AI’s current limitations in the context of architectural design.
Nitsan Bartov, architect and AI researcher at Henning Larsen, raised one of the central risks for architects: AI can generate highly convincing images of things that may be physically impossible to construct. A building designed in concrete can be transformed into timber with a simple prompt and emerge as a seductive image, even if the result makes no structural sense.
While AI might not know the difference, the architect must. In architecture, representation carries responsibility; a compelling rendering can influence clients, approvals and expectations. If AI accelerates image production without preserving the knowledge behind it, the profession risks losing credibility. This is exactly where a partnership like that between Henning Larsen and Krea earns its place. No tool removes the risk on its own — but the right collaboration can build the workflows, shared literacy and critical habits that keep creative speed and technical rigor in the same room. That’s the real work, and it’s what this partnership is set up to do.
Henning Larsen’s Head of Digital Adoption, Eliana Nigro speaks at Henning Larsen’s headquarters in Copenhagen. Photo by Martina Lanotte.
Nigro made a related point: the quality of AI output depends heavily on the person using it. When generative AI image creation first became mainstream, a generic “AI architectural style” spread quickly across social media. The designers who moved beyond that were those with clearer ideas and a nuanced approach to AI tools.
“AI becomes a megaphone for your ideas,” she said.
Human Judgment Remains Central
The final audience question was arguably the most provocative: Could AI eventually replace architects if given enough qualitative data? Rodriguez challenged the premise of the question, noting that designers and their scope of work do not fall into a single category. Bartov pushed back more directly, arguing that architectural quality is subjective, culturally specific and difficult to measure. There will always be design considerations that need human oversight.
Nigro and Bartov connected the discussion back to the foundations of architectural practice. Good architecture still requires studying history, visiting places and developing a feeling for materials and human behavior. “Feeling space, materials, empathy, ethics — these are human aspects you always stay in control of, and this is what makes good architecture,” said Nigro.
She added that, while workflow optimization is beneficial, the ability to be discerning is as critical as ever for architects. “The ability to choose well is critical, and to choose well, you need to have a lot of knowledge.”
It’s a question that deserves far more than a few minutes at the end of a panel, and one Krea and Henning Larsen will be exploring in depth in an upcoming article on Architizer.
Krea’s Co-Founder and CTO Diego Rodriguez (center) speaks with Henning Larsen’s Nitsan Bartov (left) and Eliana Nigro (right) at Henning Larsen’s headquarters in Copenhagen. Photo by Martina Lanotte.
Krea is a powerful tool for transforming design communication, but the lesson from Henning Larsen’s partnership goes further than the tool itself. Leading architecture firms can and should begin to build the cultures and critical frameworks needed to use these tools well.
AI will continue to help architects visualize faster and test ideas with greater fluidity. But as Nigro’s “megaphone” analogy suggests, the quality of the outcome will always depend on the quality of the thinking behind it.
Explore Henning Larsen’s recent projects here and learn more about how Krea Enterprise can benefit your team’s creative workflow here.
The agreement covers data security, copyright protection for AI-generated imagery, and EU AI Act compliance where applicable.
