AI strategy without the jargon bingo

You start with a conversation, not a tool. What does your business do, where is the bottleneck, where does time get lost on repetitive work, and where is the pain biggest. That is the AI Quickscan. From it come three to five use cases that are concrete candidates. You pick one to start with, small and scoped, to see if it fits your way of working. Only when that pilot delivers value do you look at the next one. We do not talk about "AI transformation", we look at one work process at a time and solve something there. You can also start with /ai-scan, a free intake call where we sketch the first outlines together.

We give no fixed promise on this, because it depends on the complexity of your organisation and how many processes are involved. We do work in phases so you do not wait months for one big report. The Quickscan phase is short and yields a first set of use cases directly. After that you do a short deep-dive per use case: process analysis, vendor comparison, ROI thinking framework, AI Act check. You can decide after each phase whether to continue. No vendor lock-in, no required annual contract. Strategy with us is not a separate end product, it runs alongside implementation.

Yes, that is exactly the SMB question. Large companies have a data department doing the inventory. At an SMB that inventory often is not there, and the owner is the one who knows everything. We do the data readiness audit for you: what data do you have, where is it, what quality, and may you legally use it for AI. Often it turns out you have more than you thought, just spread across different systems. Sometimes it also turns out you need no own data at all for the first use case and off-the-shelf AI with good prompts gets you far.

The AI Act classifies AI systems by risk: prohibited, high-risk, limited-risk, and minimal-risk. For most SMB applications (content generation, customer service bots, internal automation) you fall under limited-risk or minimal-risk, with transparency and documentation duties. High-risk applies to specific use cases: HR recruitment, credit scoring, biometrics, critical infrastructure. We do a short risk classification per use case so you know what applies to you and what documentation you need to keep. Sector-specific? For education, legal, accountancy, and HR stricter rules often apply, we have separate pages for those (see /ai-onderwijs, /ai-juridisch, /ai-accountants).

By testing it against three questions. One: does this process today demonstrably cost time or money, and is that quantifiable in hours-per-week or errors-per-month. Two: is the AI solution for this problem proven to work elsewhere, or are we the first to try it (being first is expensive). Three: are the people who have to work with it willing to adapt their way of working. If any of the three is no, we do not do it, or not yet. We use an in-house ROI thinking framework, no percentage promise. Better to say honestly that a use case saves 4 hours a week than to name a fictional number that sounds marketing-pretty.

Both, depending on what fits. Sometimes we build it ourselves because the use case calls for it and it is faster. Sometimes we coach your own team or an external builder you already have, because that is cheaper and more sustainable. Sometimes the advice is: just buy this off-the-shelf tool, it would be a waste to build it yourself. We have no interest in always building ourselves, that would undermine our advisory role. The choice between build, buy, and coach we make per use case based on what delivers most for your situation. Related services where we do build ourselves: /ai-content, /ai-agents, /ai-data, /ai-klantenservice.

We see this often. Someone signed a licence on an AI platform, there was enthusiasm, but it is not being used or it does not deliver what was promised. We then do a short audit: is it the tool, the setup, the prompt strategy, or the team adoption. Sometimes the tool is fine and only the embedding is missing. Sometimes the tool does not fit and a different choice is needed. We are not partner of a specific vendor, so we advise independently. If you are locked into a contract, we look at what is still feasible within what you already have before we propose buying something new.