Every few weeks, someone tells me that a Nordic technology company entering Southeast Asia has to "localise" before it can win. When I ask what that actually means in practice, the answer is usually some version of the same thing: be less direct in the room, be less open about price, and stop asking difficult questions so early in the conversation.
I have spent twenty-five years in enterprise technology, on both sides of the buying table, in Europe and now in Asia Pacific. I think that advice is close to backwards.
The principles that have come to be associated with Nordic business, relatively flat decision-making, transparency, long-term thinking, and a serious sense of responsibility around technology, are not a handicap in this region. They map almost exactly onto the problems Southeast Asian enterprises are running into as AI moves out of the pilot phase and into production.
The qualification matters, though, and it is the whole argument of this piece: the principle can travel even when the operating model cannot.
That distinction becomes critical with AI. Conventional enterprise software is bought, implemented, and largely settled. AI systems open questions about model behaviour, data provenance, security, human oversight, ongoing monitoring and accountability that do not close when the vendor's implementation team leaves the building. For a company buying AI, the question is no longer only whether the technology works. It is whether the organisation can govern what happens after it is deployed.
That is where Nordic business practice stops being a cultural quirk and starts being commercially interesting.
Governance has become part of the product
Nordic technology companies have generally grown up in environments where data protection, security, transparency and responsible technology practice are built into the product, not bolted on at the procurement stage to satisfy a questionnaire. That instinct is now directly relevant across Southeast Asia.
Enterprise buyers in Singapore, Thailand, Indonesia and Vietnam have moved past experimenting with generative AI. They are now confronting the practical questions that arrive with deployment:
Where is our corporate data processed, and under whose jurisdiction?
Which models are actually being used, and what changes when the vendor upgrades them?
How is information retained, and for how long?
What happens when the system produces a confidently wrong answer?
Who is accountable when it sits inside a business-critical workflow?
The maturity of those questions varies enormously by market and by organisation. Some boards ask them on day one. Others discover them the week after go-live. But the direction of travel is not in doubt, and it creates a clear opening for companies that can explain their approach in commercial rather than purely regulatory language.
No customer wants a lecture on responsible AI. They want to know what responsible deployment means for their business: what information enters the system, who can access it, what controls exist, how performance is monitored, and what happens when the system behaves outside its intended parameters.
A Nordic approach that puts those questions inside the sales and implementation process — rather than waiting for legal to raise them, becomes a differentiator instead of a constraint. In my experience it also shortens deals rather than slowing them, because the objections arrive early, while there is still room to design around them.
Flat organisations still need local operating intelligence
The same logic applies to the Nordic preference for low hierarchy, but here the translation problem is real, and I have watched people get it badly wrong.
The underlying principle is sound and highly relevant to AI: technical and operational decisions should benefit from the people closest to the problem. Models fail in edge cases. Data quality problems are usually found by operational teams, not executives. The people running the process know the workflow exceptions that will never appear in a vendor demonstration.
The difficulty is translating that principle into organisations where decision-making is more hierarchical and more relationship-oriented.
A Nordic company arriving in Thailand should not assume that encouraging employees to "speak up" will produce the same result it produces in Copenhagen. An employee may see a serious problem with an AI workflow and still not be comfortable challenging a senior executive in an open meeting. That is not an argument for abandoning open decision-making. It is an argument for building better mechanisms around it.
Structured review processes. Clear escalation channels. A named owner for AI risk. Deliberate, scheduled opportunities for technical teams to challenge the assumptions in a business case. These achieve the same objective without depending on one particular cultural style of communication.
For AI governance, that distinction is the difference between a control that exists on paper and one that works. Psychological safety is a principle. The mechanism for creating it is contextual.
Long-term thinking matters more once AI reaches production
The second Nordic characteristic that travels well is the willingness to think past the initial transaction. This matters because the economics of AI are routinely misunderstood at the point a project is approved.
A demonstration looks inexpensive. Production introduces inference costs, integration work, security controls, monitoring, model evaluation, data management, employee training, and depending on the application, permanent human oversight. A solution that looks compelling as a proof of concept can carry a very different total cost of ownership once it becomes part of an operating process.
This is why the distinction between an AI vendor and an AI partner keeps getting sharper. The vendor can demonstrate the model. The partner can explain how the system will be operated, measured, governed and supported over the next three years.
Southeast Asia makes this especially interesting because the region contains such different levels of digital maturity. A multinational in Singapore, a Thai financial institution, an Indonesian consumer platform and a mid-sized manufacturer in Vietnam may all be evaluating AI in the same quarter, and every one of them faces different infrastructure, regulatory, data and organisational constraints. There is no single regional playbook, and anyone selling one should be treated with suspicion.
A company that brings a consistent philosophy on governance and long-term value, while adapting the implementation to the local operating environment, has a far stronger proposition than one that brings a fixed methodology and hopes the market adjusts.
Transparency has to extend to the economics
There is one more Nordic characteristic that deserves more attention than it gets: commercial transparency.
AI procurement is unusually exposed to information asymmetry. Vendors understand their models, infrastructure and pricing structures far better than buyers do. Buyers understand their own operational environment far better than vendors do. Neither side has the full picture, and that imbalance is where expensive decisions are made.
Transparent pricing is part of the answer, but it goes well beyond publishing a licence fee. Enterprise buyers need visibility into usage-based costs, model dependencies, integration requirements, data limitations, and the assumptions sitting underneath a projected ROI. Change one assumption about volume or accuracy and the business case can invert.
The same principle should apply to technical performance. A credible AI partner should be able to say where its system performs well, where performance degrades, what data it depends on, and where human review remains necessary. None of that is uniquely Nordic. But it is an area where Nordic companies can convert an established business instinct into a clear market position.
In a market where most AI propositions still open with what the technology can do, there is real commercial value in starting with what it cannot reliably do. It is also, in my experience, the fastest way to earn a second meeting with a serious buyer.
The opportunity is translation, not export
None of this means Nordic companies should arrive in Southeast Asia assuming their business culture is superior. That would defeat the entire point, and the region has a long memory for people who have tried it.
The strongest Nordic companies entering this market will hold on to the principles that make them distinctive while adapting the mechanisms through which those principles are expressed. They can stay transparent without being culturally tone-deaf. They can keep flat decision-making while recognising how seniority shapes enterprise relationships. They can think long-term while accepting that relationship-building is part of the commercial process, not a delay to it. And they can take a careful approach to AI deployment without looking resistant to innovation.
That combination is going to matter more as Southeast Asian enterprises move from experimenting with AI to running it. The next phase of this market will not be won by the fastest model, the largest feature set or the most impressive demonstration. It will favour the companies that can help a customer answer four harder questions:
Can we trust this technology?
Can we govern it?
Can we justify the economics?
Will it still work for us two or three years from now?
Those are questions Nordic technology companies are unusually well positioned to answer. So the opportunity here is not to leave your values at home. It is to translate them into a form that solves the problems this region's enterprises are actually facing.
That is a far more interesting commercial proposition than cultural adaptation alone.
In our first article, we set out the 5 Power Questions Before Investing in AI — the questions leadership teams should be able to answer before any vendor conversation begins. This piece is about the standard you hold yourself to once those answers exist.
8ALTA gives boards, founders and executive teams the decision architecture to validate AI opportunities, vendors and execution plans before capital and credibility are at risk. Clarity before commitment. Start the conversation →