AI integration
AI integration means applying AI to the specific steps in your business processes where it demonstrably beats the alternative: reading documents, classifying requests, drafting responses and finding answers in internal knowledge. We integrate it with the systems you already use, with human oversight where the stakes are high and results measured against a baseline. If a simpler solution wins, we'll tell you.
Discuss your processAI is genuinely useful for a specific class of work. Language models can read and summarise documents, pull structured data from free text, classify incoming requests and draft replies. Prediction models can find patterns in historical data, such as which assets will need attention first. Used well, they take over the repetitive part of work that until recently only a person could do.
It is not a strategy, and it is not the answer to every problem. Much of what is sold as AI is better solved with a clear procedure, a fixed rule or ordinary automation, which is cheaper, more predictable and easier to maintain. We start from the process and only apply AI where it adds something the alternatives cannot.
AI for SMEs is rarely about large platforms. It is usually one well-chosen step in a process that is document-heavy, high in volume or dependent on knowledge scattered across files. Typical applications:
Where AI does not belong: decisions a fixed rule handles reliably, processes with too few cases to justify the effort, and high-stakes decisions without a person to review the outcome.
We follow the same five phases as every engagement. The question is never 'where can we use AI?' but 'where does this process lose time or accuracy, and is AI the best way to fix it?'
A working, monitored part of your process, not a demo. You get:
It can be, if it is designed that way from the start. We choose models and hosting based on how sensitive your data is, prefer business terms under which your data is not used to train models, send only what a task needs, and keep access limited and logged. Where data must stay within your own environment or within the EU, we design for that.
The GDPR applies to AI just as it does to any other processing of personal data: you need a lawful basis, should process no more than necessary, and need agreements with the providers involved. The EU AI Act takes a risk-based approach; most business process applications carry limited obligations, mainly around transparency, while uses in areas such as recruitment or credit assessment face stricter requirements. We assess which rules apply to your use case and design for them, and recommend legal advice where the situation calls for it.
AI makes sense for work involving language, documents or patterns that are hard to capture in fixed rules, and where the volume makes the effort worthwhile. It makes less sense where a simple rule, lookup table or standard automation does the job, or where errors are costly and no one can review the output. When a simpler solution wins, that is the advice you'll get.
That is not how we apply it. AI takes over the repetitive part of a task, such as reading, sorting or a first draft, so people spend their time on judgement, exceptions and contact with customers. In our radiology work with Amphia, for example, the model makes suggestions while the specialist always makes the final call.
Yes, and that is usually the point. We connect AI to the mailbox, document storage, CRM or ERP you already use, so results land where your team works instead of in a separate tool. Where a system has no usable interface, we look for a reliable alternative route and are clear about its limits.
Healthcare
AI suggestion model for radiology
Together with Amphia hospital we developed an LLM-based suggestion model for radiology: AI that supports radiologists in their assessment work with suggestions, while the specialist always makes the final call.
AI integration
Green infrastructure
Prediction model for tree condition
For Krinkels we built a prediction model that forecasts tree condition, so inspections and maintenance can be planned where they matter most instead of on fixed schedules.
AI integration · Data management
Describe the process you have in mind, and we'll give you an honest view of whether AI is the right tool for it.
Discuss your process