In 2018, Hamburg Port Authority's (HPA) CIO Lars Wentorp made a decision that most enterprise leaders still debate: put everything on one platform. Not just IT. Every department, every workflow, every employee in an organization that supports over 600,000 jobs across Germany.

"Our objective," Lars says, "is to digitize all workflows. We want to eliminate paper, both internally and in our interactions with external supply chain participants. That requires getting all departments on one platform, with common data and a map of dependencies."

Eight years later, the bet has paid off. 90% of HPA's 2,000 employees use the platform every month. In 2025 alone, they completed more than two million actions—an average of 1,000 per person. Service requests have doubled from 50,000 annually in 2020 to 100,000 today. User engagement is up 88% month-over-month. And knowledge article usage has grown 89% year-over-year.

More importantly: when AI arrived, HPA was ready.

From library to launchpad

The platform that started with IT Service Management in 2018 now runs HR, health, compliance, risk, facility management, and customer service across Europe's third-busiest container port, alongside Strategic Portfolio Management, App Engine, and IT Asset Management. The ServiceNow AI Platform operates across an organization that cannot pause—and in March 2026, it replaced SharePoint entirely as HPA's intranet, becoming the single front door for every service HPA offers, internally and externally.

"Not ten platforms, but one," says Lars. "And if you've used it once, you'll use it twice—you’ll use it forever."

The architecture that makes this possible is what IT Architect Torben Setter calls an "action-driven intranet." The traditional intranet, as Lars describes it, is passive: "a library of information—that’s where the service ends." Now, employees don't just read what the HR department does—they click a button and initiate a workflow. A manager approaching the end of an employee's probationary period doesn't need to remember to act—the platform sends them a task automatically. Risk management that used to live in spreadsheets maintained by a central team now runs through dedicated risk coordinators embedded in every business unit, each working inside the same system.

The breakthrough that changed everything

Getting to 90% adoption required something specific: making the platform genuinely useful for everyday needs, not just IT workflows.

During the COVID pandemic, HPA's health department put testing appointment booking online. Employees who had been ambivalent about IT systems discovered they preferred the platform to every alternative. "That was really our breakthrough with employees," says Lars. "They found something they needed, it worked easily, and then they didn't stop."

Then, when HPA moved payslip access onto the portal, 40% of the entire workforce showed up within days of each monthly release. They came for the payslip. They stayed to explore other services. Platform usage spiked visibly every month, creating a habit that formed around something practical.

When AI arrived, the framework was already there

In May 2025, HPA deployed Now Assist for IT Service Management. What they got wasn't a better search engine—it was AI that could use eight years of accumulated operational context to provide answers, recommend actions, and improve IT’s productivity.

Many incidents resolved by HPA's support team automatically become a knowledge article for the next person who encounters the same problem. The system analyzes patterns across thousands of incidents to identify where documentation gaps exist. A Knowledge Coach continuously reviews articles against quality standards. AI categorizes incoming requests, identifies similar historical cases, and surfaces solution strategies—reducing the cognitive load on support staff and accelerating resolution before a human has to intervene.

"Now Assist is already handling a significant volume," says Torben. "And the numbers are only partial after just one full year of coverage."

The platform's infrastructure framework was also recently strengthened. A migration to RaptorDB—ServiceNow’s real-time analytics database--reduced load times on high-use data views from approximately 60 seconds to two seconds. HPA now has the infrastructure that makes AI-scale analytics on live operational data possible: real-time queries running against eight years of compounding workflow history, without extraction, without lag, without a separate analytics environment. Torben describes a colleague who "used to go get a coffee while it loaded." Now it's instant.

What comes next

With agentic AI capabilities in development, the target is for 55% of IT requests to be handled without human intervention—a credible ambition for an organization whose AI layer is already operating inside consolidated, high-quality data. The use of automation and predictive intelligence is predicted to save tens of millions of euros over three years.

"We now want to actively exploit the capabilities of the platform for significant cost reduction and efficiency gains," says Lars.

But the number Lars comes back to isn't financial. It's 90%. An engagement rate that most enterprise platforms spend years trying to reach, achieved by a port that runs around the clock, across departments that range from occupational health to railway operations to facility management, and accessed primarily by blue-collar workers who don’t have direct access to computer systems.

"The most significant and lasting impact," he says, "has been the shift from a siloed organization to one that thinks and operates in terms of services. The platform has forced departments to look beyond their own boundaries."

Eight years ago, that was a bet. Now it's the foundation for everything that comes next.