Practical AI in your processes and your software
AI implementation
AI is on the agenda at many companies, but it often stays there: where do you start, what does it deliver, and what happens to your data? We start with your work, not with a model, and build a finished application for each process. What we propose, we build and maintain ourselves.
Start small, actually finish it
How we implement AI
First decide what good looks like
Before we build anything, we collect real examples with the people who do the work: questions, documents and outcomes that are right. That becomes the test set we measure every version against.
Choose the model on results
We do not pick a model by name. We test a few models on your examples and choose the one that scores best on them, preferably with processing inside the EU.
Build it in with control
The application goes into the software your people already use. AI output is marked, points to its source, and a colleague approves what matters. If you do not have that software yet, we build it first as custom software.
Keep measuring
Every change and every new model version goes through the test set again. So you see whether it gets better, not only when a customer notices. Sometimes an integration or a good report turns out to be enough; then we say so. You can read examples in our cases.
From one process to AI that keeps working
Discovery: where does AI add value?
We walk through your process with the people who do the work and look at the data you already have. You get a short list of opportunities: what each one delivers, which data it needs and what the risk is. With it comes our recommendation on which to do first, and why the others can wait.
Getting the data underneath in order
We bring together the data that first application needs, from systems such as your ERP, tills, accounting or telephony, while those systems carry on running as they are. You get a working integration, a dataset you can build on, and written agreements on how every figure is calculated. How long this takes depends on how many systems are involved.
Building it into your own software
We build the application into the place where your people already work, with clear boundaries: what the AI may and may not do, which data it uses and where a person approves. You get a working version in use with a real group of users, with tests, logging and monitoring in place. In the screen itself you can see what was generated by AI, what it was based on and who approved it.
Handover and building further
We agree where the source code and the application will live; by default that is a repository owned by your organisation and your own environment on Azure or AWS. You get documentation and training for your team, including colleagues without a developer background. We also agree how you measure whether it works and who picks things up when something breaks. After that you can carry on yourselves, or we keep maintaining and extending it. Are your people building further with AI themselves? With Managed AI we keep that in view and secure.
From one process to AI that keeps working
Four steps, each ending with something in your hands that you can judge for yourself. We also stop if it turns out along the way that there is little to gain. You hear that from us then, not at delivery.
Cases built with this expertise
Rhythm
Rhythm helps healthcare organisations plan capacity better. Chief works on the design of its new software suite and an AI-supported development process.
Curious what else we have built?
Frequently asked questions
How do I get started with AI in my company?
With one process, not with a plan for the whole organisation. Pick work that takes a lot of time right now or keeps being left undone, look at the data you already have on it, and build one application for it that is finished. We do that discovery with you, and we say honestly when an integration or a report would return more than AI.
Does AI work inside our existing software?
Yes, and that is usually the best place for it. We build the application into your own portal, back office or app, so your people do not have to open another tool. Taking over and extending existing software is work we do often. If there is no source code or documentation, we first map out what is actually there.
Who owns the code and the data?
We agree that up front, and the default is that it is yours: the source code in a repository owned by your organisation, the application in your environment on Azure or AWS, your data with you. We also agree which data the application may see, where it stays and how long anything is kept. So you are not tied to us to be able to carry on.
What do we need to bring ourselves?
Three things: someone who knows the process well and has time to take part, access to the systems that hold the data, and someone who can make decisions. You do not need a data department or a data warehouse of your own, we build that foundation. Messy data is not a blocker either, sorting that out is exactly what step two is for.
How quickly can you start?
When we can start building depends on what is running with us at that moment, and we tell you that in the first conversation.
What happens after delivery?
You get the code, documentation and training, so your team can use and extend the application themselves. If you would rather leave maintenance with us, that works too: we maintain software that has been running at clients for years, and we build our own products Wayzo, Spotten and Spotshare. An AI application does need maintenance: models change, your data changes, and you want to keep seeing whether the outcomes still hold up.
Want to talk through your first AI step?
Send us your question and someone who does this work reads it, not an account manager. In a first conversation we take one process apart: what is possible, which data you already have, and what the smallest version is that returns something. You will also hear it from us when the answer is that you should build something else.
Monday to Friday
8am to 5pm
Send us a message
What else can Chief help you with?
Do you need new software built, or existing software taken over and developed further? Or are you looking for senior developers and technical leadership for your own team? Take a look at our other services.
Custom software
Custom software development from Utrecht: platforms, apps, integrations, hardware and firmware, from business case to support. See how we work.
Developers & technical leadership
Senior developers and technical leadership from one team in Utrecht. You hire a developer, a tech lead, or both.
