26 JAN 2026

From data to debate: Mapping Norway’s AI ecosystem with RankmyAI

Written by Wilco Verdoold

An interview with Bram Timmermans, professor of innovation & entrepreneurship and research director at Digital Innovation for Sustainable Growth.

When you try to map 'the AI industry', you quickly encounter an awkward truth: there isn’t one. Not as a neat sector, not as a clean category, not as a box you can tick in standard classifications. AI is everywhere and nowhere at once, embedded in products, woven into services, and tucked into processes. That makes a simple question surprisingly hard to answer: who is actually building and using AI, and how do we know?

To unpack that challenge (and why it matters for both research and policy), I spoke with Bram Timmermans, who has spent the last two decades in Scandinavia and now works at the Norwegian School of Economics. Alongside his role as professor, he is research director at Digital Innovation for Sustainable Growth, a centre that collaborates closely with both private and public organisations in Norway. Together, we discussed Norway’s AI adoption, the problem of 'AI washing' , and why rankings are most valuable when treated as a starting point for dialogue, not the final word.

Norway adopts fast but rarely builds the foundational tech

Norway is often described as digitally mature: strong connectivity, widespread digital banking, and long-standing expertise in areas such as mobile communication. Timmermans sees a low barrier to adopting digital technologies, including AI.

But he adds an important nuance. Norway is primarily an advanced user of AI, not typically a place where the most fundamental technology is developed. Like many smaller countries, it often draws strength from combining AI with deep domain expertise.

That combination shows up most clearly in areas such as:

Energy
Fisheries and maritime
Finance

Adoption is not evenly distributed, especially among SMEs, because capabilities and resources vary widely once you zoom in beyond the national headline numbers.

The classification problem: AI doesn’t behave like a 'sector'

A core issue Timmermans keeps returning to: “You have no industry class that’s called artificial intelligence, because it’s integrated.” AI is absorbed into organisations that still look like energy, logistics, finance, or manufacturing in every formal dataset. That becomes a real problem for anyone trying to understand an AI ecosystem, whether you’re a researcher, policymaker, investor, journalist, or a business looking for partners. If the category is fuzzy, your list of 'AI companies' will always be incomplete or distorted. Not because people are sloppy, but because the underlying structure is messy.

AI intensity scores, manual checks, and the reality of AI washing

Before working with RankmyAI, Timmermans and colleagues explored approaches such as scanning company websites and creating 'AI intensity' signals. In practice, this involves assessing how strongly companies present themselves as AI-related. It can work, but it quickly becomes heavy. You still need manual validation, because mentioning AI does not mean you are using AI meaningfully. Companies have learned that the word AI is good marketing. Timmermans refers to this as AI washing, the AI-era cousin of greenwashing. A company might mention AI because it sounds modern, because they once ran a pilot, or because it helps in sales and recruitment. That does not automatically reflect real capability or real value creation.\

Why RankmyAI: a dataset that creates conversation (not closure)

So why collaborate with RankmyAI?
Because it offers a structured way to identify and compare AI-related companies using consistent signals, while being open about its limits, Timmermans sees transparency as important; he is explicit that the ranking should not be treated as a definitive scoreboard.

“We always say: the ranking is a dialogue tool.”

He also emphasises the importance of context. RankmyAI’s approach stems from a background in which marketing and digital commerce signals matter, shaping what the ranking is sensitive to. That is not “wrong,” but it does mean you should interpret results through a local lens and improve them with local knowledge.

In Norway, there is also a practical advantage: the ecosystem is large enough to matter, yet not so large that it becomes impossible to validate. If the list contains a few hundred companies, you can still perform meaningful manual curation, spot mismatches, add missing players, and handle edge cases.

From visibility to policy: short lines, high trust, real access

One striking part of the conversation is Timmermans’s description of the policy environment in Norway.

Norway is a small country with relatively close institutional networks, and he describes it as a high-trust society with shorter lines between academia, industry, and government. That makes it more normal to engage directly with policymakers, sometimes even at the ministerial level, especially when you’re working from a well-connected institution and a research centre that actively collaborates with stakeholders.

Has the AI ecosystem mapping already changed policy?

Not yet, Timmermans says, partly because it’s still early. But it has helped ignite discussion, and that’s the first step: creating shared language and visibility around what’s happening, what’s missing, and what strategic decisions may be needed.

He also notes the timing: the Norwegian AI report and its public discussion landed around an election period, which can both complicate and amplify the debate.

What’s next: the “visibility” that matters for ecosystems

When asked what RankmyAI could add to become even more useful, Timmermans points to a form of visibility that goes beyond website traffic and reviews.

Media presence and narrative signals
How often is a company mentioned in the media? Not as consumer hype, but as ecosystem relevance. Particularly useful in B2B contexts where traditional “consumer signals” can be misleading.
Network connections
What partnerships, references, and ecosystem links surround a company? Early-stage companies may be “important” because they’re visible in the ecosystem, not because they already have massive revenue or mainstream adoption.
Jobs and hiring signals
Open roles (and the types of roles) can reveal momentum and strategic direction. They’re a strong proxy for growth and capability-building.

Importantly, he recognises a constraint: these signals are often easier to gather per country than globally. That’s where partnerships become essential. Local researchers and institutions can help enrich and validate national datasets in ways that are difficult to scale worldwide.

Beyond borders: why Nordic comparisons matter (and why use-cases are hard)

Timmermans also zooms out to the Nordic level. The Nordics are often treated as a single block in international narratives, but he stresses they are not a single entity. There are meaningful national differences in technology, innovation structures, and industrial focus. A geography-based approach provides a practical boundary (this country, that country). But the more interesting ambition, especially for global industries, is to map AI by technology and use-case, regardless of borders. That’s also much harder.
Once you remove geography as the organising principle, you need new ways to define and validate:

What counts as a use-case,
how companies cluster around it,
and which signals indicate genuine adoption rather than marketing-speak.

The takeaway: rankings are valuable when they provoke better questions

If you use a ranking as a structured entry point to trigger debate, identify gaps, and enrich understanding, it is highly useful for research and policy. Or, as Timmermans effectively frames it: the ranking isn’t the answer. It’s a tool for starting the right conversation, grounded in data, sharpened by context, and improved through collaboration.