The world around us is speeding up. AI is accelerating product cycles, and customers expect answers in hours.
I feel that acceleration in my own work and in our organization. AI lets our teams run several agents at once, but it does not necessarily feel like we have less to do.
Give people more capability, and the instinct is to ask for more, not less. That puts every leader in front of a real decision. Where is speed good? Where is it excellent? Where should we expect ten times the pace, and where do we need to slow down?
Two recent projects made that distinction clear. One shipped in 24 hours. The other took a year. Both moved at the right speed. But there is another kind of speed that matters just as much. When something goes wrong, customers should not have to wait to hear about it.
That is the shift I keep coming back to. As execution gets faster, judgment gets more valuable.
AI is changing what a small team can ship in a day. 2026 has been a transformative year for how we develop code and use agents, and engineering productivity has already increased by nearly 20 percent, which I see as early evidence, not a ceiling.
That figure matters only if it helps solve the right customer problems. Feedback from technical support and customer-facing teams must shape what we decide to build. AI can accelerate execution, but it cannot decide which problem matters most.
The clearest case for speed is when the thinking is already done. One customer asked us to change the way a particular button generated insights. The request came through technical support, was clearly described, and included a proposed solution. An AI agent helped code the change, build the interface and test it. The update went live 24 hours later.
The customer need was clear, the impact was understood, and the decision about what to build had already been made. In that situation, speed is service.
Not every customer need comes with a clear answer. One example is our work to build remote firmware update capabilities with an enterprise customer.
Today, updating the software on a machine means sending a technician out to it with a laptop. Doing that remotely is not the hard part. The hard part is that no two manufacturers approve a release the same way, or define a finished state the same way.
So much of the year went into the agreement rather than the code. Who signs off. What counts as safe before an update runs. What happens when one fails.
AI can compress the build. It can’t speed up the trust.
There is a third kind of speed, and it does not require any judgment at all. When something goes wrong, the fix takes as long as it takes. Telling people should not.
I believe bad news has to travel extremely fast. When a customer-facing capability has an issue, the technical fault is only part of the problem. The other part is uncertainty. Does anyone know? Is somebody taking ownership? What happens next?
We flag incidents on our website as soon as we understand their impact. Customers should hear relevant information from us first. Perfect systems do not exist, but clear ownership and timely communication build confidence when problems occur.
AI will expose where an organization can move quickly and where it cannot. Sometimes the constraint is not the team. It is the system: technical debt, poorly defined data or an architecture that was not built for the speed the business now needs.
That exposure is useful. When execution was slow everywhere, a slow system was easy to miss. Now it stands out, and pushing people harder will not fix it. The work is to remove the constraint while staying close to customers and being clear about what is happening in the meantime.
So the job is changing. When execution was expensive, leading was mostly about getting things done. Now that it is cheap, leading is mostly about deciding what deserves doing, and at what pace.
That’s what finding the right speed means to me.
Søren Brogaard is CEO of Trackunit, a global technology company on a mission to eliminate downtime in construction.
Under his leadership, Trackunit has evolved into an AI-powered platform business connecting more than 6.5 million assets. He became CEO in 2020 after serving as CTO and CCO, following senior leadership roles at Schneider Electric. An active voice in the construction industry and media, Brogaard champions collaboration and mindful leadership grounded in purpose, trust and personal development.
