The Nordic AI Implementation Model, or NAIM, is the framework we use for every project. It has four phases: Discovery, Implementation, Enablement, and Evolution. The model addresses a specific problem: most AI projects go straight from "idea" to "build", skip discovery and enablement, and end up as a system nobody uses.
In this article I will take you through each phase, what actually happens, and what deliverables to expect.
Phase 1: Discovery (1 to 3 weeks)
The goal of Discovery is not to plan the solution. The goal is to decide whether the solution should be built at all.
In a typical Discovery we map: which workflows are candidates for automation, which data flows exist today, which systems need to be integrated against, which operations staff will be affected, and which compliance requirements affect the solution (GDPR, EU AI Act, sector-specific requirements).
The deliverable is a written document with: problem definition, recommended solution architecture (or recommendation not to build), estimated timeline and cost, and a clear risk analysis. Discovery is delivered at a fixed price, and you own the document regardless of whether we continue.
Phase 2: Implementation (4 to 12 weeks, typically)
Implementation is the build. If Discovery concludes with a go, this is where we start.
For Own deliveries everything is built in the client's environment. We set up infrastructure, integrate with existing systems, configure agents or the operational layer, and bake quality and audit trails into the architecture from day one. For Lease deliveries the client is connected to the tailored version for their vertical, with the necessary integrations.
Implementation is the phase that separates a consulting delivery from a project that never finishes. We work in short iterations with weekly demos, not in 6-month silent periods with a big reveal at the end.
Phase 3: Enablement (2 to 6 weeks)
This is where the real value sits. Enablement is how we make sure the solution actually gets used.
It includes: practical training for operations staff, documentation that actually gets read (not a 200-page handbook gathering dust), establishing a support channel into Nordic AI, and measurable KPIs that track adoption and value.
Most AI projects skip enablement or do it half-heartedly. The result is always the same: the system exists but is not used. That is why we make enablement an explicit phase with its own deliverables, not a "we will throw some slides at the end".
Phase 4: Evolution (ongoing)
AI systems are not "done" the way a regular software project can be done. Models become outdated, workflows change, new requirements emerge. Evolution is the phase where Nordic AI continues to monitor, adjust, and develop the solution further.
For Own deliveries this is offered as a monthly maintenance package or retainer. For Lease it is included. The client can always choose to take over Evolution themselves once their team has grown into the role.
What this means for you
If you are considering an AI investment, the first step should be to schedule a Discovery, not to buy a solution. If you have already bought a solution that is not being used, chances are the enablement phase was skipped and that is where the effort should be moved.
We offer Discovery as a standalone delivery. Get in touch if you want to talk through your project.