On a recent weekend, I built a version of the classic video game Pong for my kids using AI.
It’s a simple example, but it shows how quickly the gap between idea and execution is disappearing.
At Nokia, we’re seeing the same compression in our work. It means engineering teams can explore multiple architectural paths in parallel, test them quickly and achieve stronger outcomes faster. Work that was once categorized as operationally difficult or too time-consuming becomes achievable – opening up new opportunities while reinforcing the importance of focus.
More output, faster delivery
The real productivity gain is more output from the same teams, delivered faster for customers. That happens when AI moves from individual use into repeatable workflows.
Since the broad rollout of the AI coding tool Cursor across Nokia earlier this year, more than 14,000 of our team are using the platform across software R&D, with weekly active usage at 67% – and growing. Six months ago, much of this was experimentation. Today, we are seeing repeatable patterns emerge.
In one engineering workflow, teams using AI-assisted development compressed a four-month feature timeline into a couple of weeks. In another, system-level test cases that previously took hours or days to build can now be created in minutes.
The improvement comes from connecting intent, context and implementation more directly. The best results flow from teams iterating faster with greater consistency.
More judgment, less coordination
As execution gets faster, the constraint inside organizations moves upstream.
Historically, scaling output often meant adding coordination, layers and process. In an AI-enabled environment, that model breaks. Scaling up requires quicker decisions and teams with more autonomy to act.
The teams seeing the biggest gains so far have combined AI with deep domain knowledge, engineering discipline and clear guardrails. We have found the greatest value is created by people who understand the problem, own it end to end, and apply the technology with judgment.
AI tools do not replace expertise. They increase the leverage of expertise.
Technical judgment, customer judgment and business judgment become the differentiators.
That also changes what a high-performance team looks like. A “star team” is not built only by recruiting for capability against the work. It is built by creating a team comfortable operating with ambiguity, less coordination and continuous learning, individually and together.
One of our engineers described the impact of AI as: “lowering the cost of curiosity.”
I think that captures it perfectly. Lowering the barrier to testing ideas means more options can be explored before a team commits. It also changes who can contribute and increases the importance of people with different ways of thinking.
Ultimately, for curiosity to create consistent value for customers, that requires a high-performance culture and a different approach from leaders.
Leadership at a higher clock speed
In a recent town hall, I was asked by one of the team: “How are you using AI in your day-to-day work?”
I use AI to prepare for meetings, explore technical questions, review material and support how we operate as a leadership team. In many cases, I now use it more than traditional search.
What was most interesting about the question was what it revealed about adoption. People are not waiting for AI to be cascaded through the organization. They are increasing AI usage in their own work, and expect leaders to do the same.
Moving at pace is not enough. Leaders need to define priorities, make trade-offs and create the conditions for teams to operate at a higher clock speed.
They also need to be closer to the work. Not to micromanage, but to understand how teams are using these capabilities, where bottlenecks remain and how decisions move through the organization.
According to McKinsey, nearly 90% of organizations are using AI in at least one part of their business, yet only around one-third have scaled it across the enterprise.
At Nokia, how we work is connected to what we build.
As AI workloads move beyond the data center, networks need to do more than carry traffic. They need to deliver AI tokens for the task at hand. That is a structural shift in customer expectations that requires a fundamentally different network architecture.
It is not enough to layer intelligence on top of networks. Networks must become AI-native by design.
The same principle applies inside companies. The companies seeing the greatest impact are not just adopting AI tools. They are redesigning how execution happens.
We see a similar shift happening across AI infrastructure.
To serve customer demand, it is not enough to simply layer intelligence on top of existing systems. Infrastructure needs to become AI-native by design.
The opinions expressed in Fortune.com commentary pieces are solely the views of their authors and do not necessarily reflect the opinions and beliefs of Fortune.
