From systems of record to systems of action: the new reality of AI

Developments in artificial intelligence are happening faster than ever. For years, IT systems primarily recorded information and supported processes, but we are now seeing the rise of “systems of action”: systems that not only store data but also actively use it to take work off your hands. This presents a unique opportunity for a vertical market software company like Topicus, says Egbert van der Veen, AI lead within the Education division.

The role of AI within Topicus

"AI now touches everything: our products, our processes, and our customers," says Egbert. As AI lead, he connects initiatives within Topicus. "We're a company with extensive domain expertise. AI projects are already underway in various divisions, but the real value comes from learning from each other. I make sure people connect, that we recognize overlapping challenges, and that we deploy the right technology in the right place."

Egbert also translates global technological developments into practice within Topicus. "You need someone who sees what's happening outside and understands how our company works. That's the only way to connect the right opportunities with the right teams."

AI as an amplifier of the organic growth strategy

For years, Topicus has distinguished itself through organic growth: continuously building new, sustainable value within existing sectors. AI accelerates this ability.

"Where there used to be a divide between digital systems and the analog world, we can now truly bridge that gap for the first time. This means new use cases, new ways to solve customer problems, and new forms of value creation."

European technology as a foundation

While the global market is dominated by large players, Topicus is exploring European, open-source AI technology. "We work with models developed and hosted within Europe. Data cannot be transferred to Big Tech. That's essential for us and our clients. The engines we use are secure, transparent, and fit our domain."

Training your own models? Only where it truly adds value

According to Egbert, organizations should critically consider developing large language models themselves.

"It's like trying to build your own operating system when Microsoft already exists. Only in very specific scenarios, such as those involving Dutch language or culture, can fine-tuning a small, existing model be useful. But building complete large models? It's neither realistic nor necessary, given the costs involved and the fact that there are good open-source alternatives running on European hosting partners."

The value of AI for end users

At Topicus, AI never starts with technology, but with customer needs. "You have to be careful that AI doesn't become a gimmick," says Egbert. "That's why we start with customer value."

A concrete example of using AI to address challenges in education, such as lack of time due to teacher shortages and large classes, is:

• The Learnbeat grading assistant that helps teachers with open-ended questions, with the teacher as a 'human-in-the-loop'.

• Summaries of student records that can be used for ten-minute interviews, giving teachers faster and more complete insight. The teacher or mentor always takes the lead in forming an expert opinion. Human interaction and professionalism are essential in this process.

• Real-time translation of parent communication, so that parents who do not (fully) master the Dutch language can be more involved.

Responsible AI use: the five Topicus pillars

To deploy AI safely and responsibly, Topicus uses five fixed preconditions:

• Data control: know where data is and how it is managed, in accordance with the GDPR.

• Traceability: traceability of AI claims back to the source.

• Human-in-the-loop: the human always makes the final choice.

• Quality assurance: minimizing risks of incorrect output.

• Scalability: being able to roll out solutions sustainably and efficiently.

“By establishing these pillars at the beginning of each conversation, you create trust and space to look further.”

Systems of record vs. systems of action

Egbert sees a clear shift in how we use software. “Our current systems remain necessary; they form the reliable data backbone of our sectors. But AI makes it possible to build a new layer on top of that: systems that not only support, but actually take work off our hands and optimize processes. The value shifts towards action. The possibilities for saving time and reducing workload now go far beyond anything that was previously possible. For Topicus, this isn't a matter of following the trend, but an opportunity to use our domain knowledge and passion for education to actively shape this new generation of systems.”

From science fiction to reality

AI is the biggest technological development I've seen in my career. The fact that computers understand human language has long been science fiction. It's such a huge opportunity, and that's why everyone should be working on it.

According to Egbert, organizations and professionals need to overcome the chicken-and-egg problem: "You have no experience with AI, so you don't start, and therefore you learn nothing. The only way is to just start. Take small steps and experiment responsibly. Increase your productivity. From there, knowledge will grow naturally."

More information?

Please contact:

Bianca van Bel

mailbianca.van.bel@topicus.nl