TNO has conducted an exploratory study into strategies for scaling AI applications across government organisations. The study was commissioned by the Ministry of the Interior and Kingdom Relations (BZK) because AI offers numerous opportunities for the public sector.
Taking the lead on AI scaling
The Netherlands’ Digitalisation Strategy (NDS) commits the government to a more proactive role in AI. This includes exploring the establishment of an AI scaling facility to enable the reuse and further development of successful AI applications, making them deployable more quickly and more widely. A shared vision for AI scaling is essential to this process. The TNO study makes a significant contribution to this effort and forms the foundation for the ‘AI Scaling Guidance’ currently under development.
7 AI scaling strategies
The research shows that scaling is neither a one-size-fits-all nor a linear process. Depending on their goals, context, and collaborations, organisations, consciously or unconsciously, adopt different scaling approaches. The report outlines 7 AI scaling strategies, each with key considerations, to guide government organisations through their scaling journeys. As the report is available only in Dutch we summarise the 7 AI scaling strategies:
- Scaling in: Expanding an AI solution within the same organisation by rolling it out to additional departments, teams, processes, or use cases to increase internal adoption and value creation.
- Scaling out: Replicating and adopting an AI solution across similar organisations, transferring knowledge, tools, and lessons learned for broader benefit.
- Scaling beyond: Applying an AI solution in new sectors, domains, or policy areas, adapting it to different users, objectives, and operational environments.
- Scaling together: Collaboratively developing and implementing AI solutions across multiple organisations, sharing resources, expertise, governance, and responsibilities for broader applicability.
- Scaling down: Tailoring a broadly applicable AI solution to fit local circumstances, organisational needs, legal requirements, or specific user groups.
- Scaling up: Influencing policies, regulations, standards, and institutional frameworks to enable wider AI adoption and large-scale implementation.
- Scaling deep: Embedding AI-enabled innovation into organisational culture, routines, capabilities, and values to achieve lasting behavioural and cultural change.
Conversation starter
To support organisations in achieving their AI scaling objectives, a ‘Conversation starter’ has been developed. It includes questions to help organisations take stock at the outset of an AI scaling initiative. The document compiles practical recommendations, drawn from real-world experience, to help shape and steer the scaling process. Below is a brief summary of these key recommendations:
Key recommendations
- Treat scaling as a strategic choice, not as a default outcome of a successful pilot.
- Select the scaling strategy that best matches the intended impact and context.
- Focus on organisational, governance, and cultural factors as much as on technology.
- Collaborate and reuse existing solutions where possible to avoid duplicating effort across government organisations.
- Invest in AI literacy, documentation, and knowledge sharing to support sustainable adoption.
- Establish clear ownership, responsibilities, and funding arrangements before scaling.
- Continuously assess whether scaling remains desirable, feasible, and legally appropriate in each new context.
More information
For the original announcement and to read the full report (in Dutch), visit TNO’s website.
