AI is everywhere. Almost every company uses it in some shape or form. Someone summarises a meeting, someone drafts an email, someone asks it to fix a spreadsheet formula.

Still, Gartner expects more than 40% of agentic AI projects to be cancelled by the end of 2027. In Belgium, only 6% of technology companies say AI is an integral part of how they run, and 85% see no effect on turnover.

So where do these projects go wrong? In our experience, it’s almost never about the AI.

Sources: Gartner, 2025 · Agoria, 2026

AI alone is not enough

When we first sit down with a client, we often hear things like “The AI gives answers that sound right, but aren’t.”, “We ran a pilot. The demo was great, and nobody uses it.”, “It’s one more screen to check, not less work.” or “Every department does it differently.”

In our experience, the underlying issue is not necessarily the technology itself, but that agentic solutions are not built on actual business context or for the people that have to work with them.

• AI agents don’t know your business. Different systems, teams and branches often have their own definition of a customer, an order or a process. What counts as an “active customer” in office A may not count in office B. Agents then start to guess which source is the truth.

• The work stays the same. AI gets added on top of the existing process instead of changing it. It becomes one more screen to check, not less work. Or worse, it doesn’t fit the context in which it’s supposed to be used.

• People don’t trust it. When an agent can’t show where its answer comes from, or gets it wrong once too often, people stop relying on it. They have to spend time double-checking or start working around it.

At Cnext, we design for all three from day one. The foundation: a shared understanding of your business, so agents work from the right data and give the right answers. The process: where the agent fits in the work, so the work actually changes. And the people: what they need to fit it into their day and trust what it suggests. That’s what sets us apart: most providers stop at the technology, while we are one partner from first idea to daily use, so AI has a real impact on how your business runs.

AI that knows your business.

Before we build anything, we map how your business actually works: the terms you use, the systems that hold them, the rules that apply and the people who decide. In technical terms, this is our Kairos Ontology. It is the foundation of Kairos, our platform where people and AI agents work and decide together. In practice, it’s shared business knowledge, and the foundation our AI agents are built on.

It’s why an agent is triggered at the right moment, identifies the next step, prepares it and forwards it to the right person to approve. And because this knowledge is shared, the second and third use case build on it instead of starting from zero.

Our second logistics client just went live with about 75% less effort than the first. We will share their story during our webinar on 25 November. Registere here, attendance is free.

AI changes more than tooling

AI isn’t just another tool. When an agent takes over part of someone’s tasks, it changes their role, how they see themselves at work, and in the bigger picture it can change the company culture.

That’s why we care about involving the people who do the work from the very beginning. We understand how they work today, where an agent fits into their workflows, and where it doesn’t. We ask what they need to adapt and to adopt: new skills, clear responsibilities, the confidence to rely on what the agent prepares. And we stay until it’s part of everyday work, not just until go-live.

Designed for how people build trust and decide

Whether people rely on an AI agent depends on design choices: what the agent shows, how it explains itself, when it asks for a decision. Our behavioural science team draws on research and clinical practice about how people build trust, and how they lose it. We turn that into concrete design rules for every solution we build and we test it with the people who will use it, and their feedback shapes the next version before anything goes live.

We also work with ULTRA, a research project of Odisee and KU Leuven that studies the design factors shaping how people and AI work together. We bring real use cases from our projects. They test them with methods like usability studies and eye tracking. And the findings feed straight back into how we design our agents.

Where does your AI stand?

Whether you’re thinking about bringing AI to your business or are already working with it, take a moment to think about whether shared business knowledge, change management, and design principles are part of your solutions yet.

If not, Cnext is the right partner to get you there: making sure AI is built for your business and the people in it. We’ve been making technology work for businesses since 2007, across more than 300 solutions.

Curious to learn more? Come talk to us at Web Summit in Lisbon on 11 November (Pavillon 3, booth G1-33), join our logistics webinar on 25 November, or book a meeting with our team.