Adrian Sjøholt is a Solution Architect at ProsessPilotene. He has worked on implementing AI solutions on the Dynamics 365 platform for many of our customers, and he is among the first in the world to earn Microsoft’s Agentic AI Business Solution Architect certification.

Here, Adrian answers five questions that many leaders ask themselves—and us:

1) Why is everyone talking about AI agents these days? What's going on?

We are in the midst of a shift from AI that answers questions (generative AI) to AI that performs tasks for us (agentic AI). Over the past year, the technology has evolved from simple chatbots into systems that can plan, reason, and use tools entirely on their own. Traditional AI requires a human to guide it, meaning prompting at every step. An agent, on the other hand, is given a goal—for example, "Plan and carry out the entire customer journey for this complaint." The agent determines which steps need to be taken.

This marks a shift in how we interact with systems. We are moving from using tools to creating digital employees with their own profile pictures and email addresses.

2) Trends: Where Is the Development Heading?

Over the next few years, many Norwegian organizations will move from isolated AI pilots to more integrated solutions involving multiple AI agents working together. Many have already started using AI, but not all are successful—often because the initiatives lack a clear business rationale, not because the technology is immature.

At the same time, we see a clear trend towards more autonomous systems. However, the limit of autonomy lies with decisions involving high risk and legal responsibility.

"Success is 20% about technology and 80% about organization and data." – Adrian Sjøholt, Solutions Architect at ProsessPilotene

Process design is also changing. Instead of linear workflows, teams of specialized AI agents are now being built to collaborate, much like employees with different roles and areas of expertise.

As we gain experience, we are also moving towards a human-on-the-loop approach, where people increasingly monitor and approve results rather than performing each individual step.

3. What needs to be in place to succeed with AI—beyond the technology itself?

Success with AI in an organization is 20% about technology and 80% about organization and data.

4. Why are AI agents more about organizational development than technology?

The role of the solution architect is shifting from building static solutions to becoming an "agent choreographer". We will design how people and agents can work together most effectively to create business value. Going forward, leaders must understand that AI agents are not an IT project, but a strategic organizational initiative. It is about building capacity. Many people today feel that they have too much to do, and AI agents can help increase this capacity. Those who adopt them will gain a head start.

5. The expert’s advice for getting started with AI agents

If Adrian were to give one piece of advice to organizations that want to succeed with AI agents, it would be this: Think big, start small, scale fast! Choose a clearly defined problem in the organization, build an agent that solves the problem 100%—and use this success to build trust and understanding for further deployment.

▶️ Video: Competitive advantage with AI agents (April 2025)

Do you want to optimize workflows?

We specialize in automating and optimizing workflows. Contact us for a no-obligation chat.