Companies we've built for
It opens the email that just came in, writes the quote inside it into your table, corrects the wrong address, tells the customer. One of the jobs your team repeats every day, off their desk.
Your e-commerce panel, your ERP, your inbox, your calendar, your WhatsApp. Whichever you use, it connects to that one. It takes the data from one, does what needs doing, writes the result into the next. What somebody copies from one system to another by hand today simply arrives.
When it has to, it writes to your customer or to your team. It asks for whatever is missing. Most of what we build finishes quietly in the background, without talking to anyone.
It reads from your data and nowhere else. Ask it something it doesn't know and it says so, then hands the question to your team. On the critical steps it waits for your approval: orders, returns, prices. You always have the final say.
It connects to the tools you already use
From e-commerce to appointments, from automations to agents. Every one of them runs inside a real customer's day.
It checks the address and the postal code on every order, finds the ones that do not match and corrects them. No parcel goes to the wrong door, no return costs, no waiting customer. It all runs in the background.
A pricing agent for freight forwarders. It reads the request that lands in your inbox, matches route and rates, then drafts and sends the quote.
Any industry. We walk through your process with you, then build the agents and automations around it. Every job we have delivered so far started this way.
The system only reaches the data a job needs. If you want records kept, we agree at setup what is stored and for how long, and we put it in writing. If you don't, we process the data on our own infrastructure, send you the result and keep no copy. Nothing is written to disk, logged or backed up.
The frame is the same on the AI provider's side: your data never trains a model. It is held for at most 30 days for security review, then deleted automatically. As a company based in Amsterdam we fall under GDPR directly. We share the contract and the data processing details in writing before we build anything.
We ask two things of you: tell us the process, and give us access to your systems. You buy no new software and you move no data. The build, the testing and everything after are ours.
We go through your process together and pick the jobs the agent will take over.
For each job we decide: agent or automation, and where your approval belongs.
We develop the agents and automations, then connect them to your systems.
We run real work through a narrow scope and tune the details.
We open it at full scope. Watching it and improving it stay with us.
No big project invoice lands on your desk. One amount every month, like the salary of a new hire. You know what it is, in writing, before we build anything.
There is no separate invoice for discovery, design or the build. All of that is ours.
A single monthly amount. Running the system, watching it, improving it and answering your questions all sit inside it.
You get the monthly amount in writing before setup, and it stays the same as long as the scope does.
If it isn't, write to us and we answer within 24 hours. When we don't know something, we say so plainly. We taught our agents the same habit.
Ask for a live demoIt connects to the systems you already use and takes over the work your team does by hand. It moves data from one place to another, checks it, applies the rule and reports the result. When the job calls for it, the agent writes to your customer or your team, and at a critical step it asks you first. Which work it takes over, we decide together, based on your process.
A ready-made tool doesn't know your business. W Agent runs on your data, your processes and your rules, and it writes to your customer in your own voice. We do the build and the settings. You don't set anything up.
Right now only for freight forwarders: a pricing agent. In other industries we map the process with you first, then build the agents and automations for your case. Every project we've delivered so far started this way.
Not unless you want it to. By default, steps like an order, a return, a price or a record written to your ERP land in front of you for approval first. During the build we decide together which steps go to approval and which ones run straight through. Some of our customers want certain actions to run without a stop, and that is possible too. On top of that, sensitive steps such as price calculation, sending email and scheduling follow your rules exactly. There is no guessing.
If something unexpected comes up, the agent stops the work and tells your team. It won't quietly do the wrong thing. We watch the system, so most of the time we see the problem before you do. For critical steps we agree on the way back during the build.
No. Your e-commerce panel, ERP, email, calendar, WhatsApp: whichever one you use, we connect to it, read the data there and write it back there. We won't make you install a new system. You get a separate panel to follow the process and step in when you want, and your daily work carries on in the tools you have today.
In most cases, yes. If there is an API, we connect straight to it. If there isn't, we use whatever is at hand: the database, a file transfer, an email export. If there is no way in at all, we say so in Discovery and we don't start the build.
Discovery, Design, Build, Pilot and Live. Usually live in 4–6 weeks. In industries where the agent comes ready-made, most of that time goes into the settings that are specific to you.
The build is on us. We make no "do it yourself" promise. Teaching an agent the work takes expertise, and it is included in the monthly payment. You explain the process, the rest is our job.
We do. We watch the system, see what is slipping and fix it. That is what the monthly payment pays for. You don't need a separate technical team on your side.
You tell us and we update it. A new supplier, a changed approval step or a different campaign, and the build changes with it. Changes inside the current scope are included in the monthly payment.
We decide what to measure together, before the build starts: messages answered, wrong orders prevented, appointments confirmed, transactions taken off your hands. The Pilot is the first test of that. We report the numbers.
Any time. You can take the work over, stop the process or reject an approval. The system doesn't decide in your place. It applies the rules you set.
No setup fee. You pay a single amount every month. Running the system, watching it, improving it and support are all inside that amount. We give you the figure in writing before the build, and we don't change it as long as the scope doesn't change.
If a new agent, a new process or a new system comes in, the monthly amount changes too. We never do that unless you ask for it. First we agree the scope, then the figure, both in writing. Maintaining and improving the current build costs you nothing extra.
You choose how data is stored. If you want records kept, we decide together during the build what is kept and for how long, and we put it in writing. If you choose minimum retention, your data is processed on our infrastructure, the result comes back to you and no copy is left behind. Nothing is written to disk, logged or backed up. On the AI provider's side, data is held for at most 30 days for security review, then deleted automatically. For stricter requirements we turn on enterprise options such as a zero-retention agreement and data residency together with you. In every case the system reaches only the data the work requires. We share the contract and the data processing details in writing before the build.
No. We always work through the enterprise API. Anthropic and OpenAI do not use data sent through that channel to train their own models, and that is a written commitment in those providers' commercial agreements. The systems we build are not trained on your data either. For builds where records are kept, we set out in the contract how the data is stored and used, the way you want it.
We use whatever the work calls for. We are not tied to a single model. During the build we tell you in writing which provider runs in which region. If your compliance team wants that list before the contract, we give it to them.
For our customers in Turkey we work within the KVKK framework. Our operation is built to GDPR, and the two sets of rules overlap to a large degree: legal basis, data minimization, purpose and storage limitation, and the rights of the individual. We define our obligations as a data processor, and the legal basis for any transfer abroad, in writing in the contract.
Yes. GDPR is the EU data protection regulation that came into force in 2018. It binds every organization that processes the data of people in the EU, wherever that organization sits. As a B.V. based in Amsterdam, we are directly in scope. Its six principles frame the way we work:
