AI-Powered Custom Software
For SMBs that want to get one process really right. Not another SaaS subscription, not a chatbot on your website. An application you own, built around the process that runs your business, with AI doing the real work. A compact team from our builders, SMB scope with enterprise discipline.
Who it is for
Dutch SMB founders and operations leads who need an application for one specific process, with AI built in from the start. Typical client team size: 10 to 100 people.
- SMB founders
- Operations leads in growing companies
- Organisations without an in-house engineering team
Problem it solves
SMBs rarely have an in-house engineering team. Off-the-shelf SaaS does not fit the specific process that runs the business. Cobbling together Zapier, a chatbot, and a spreadsheet ages badly. What they need is a small, owned application with AI doing the work that would otherwise require another hire.
- "We spend half our week on quoting. We need a tool that writes the first draft."
- "Our intake process is email. It should be a structured form with an AI that turns emails into records."
- "We want our own tool, not another SaaS subscription, and we want AI in it."
What a typical engagement looks like
| Phase | Duration | Output |
|---|---|---|
| Scoping | 1 week | One-page brief: one process, one application, one clear outcome |
| Design and build | 6 to 12 weeks | Working application deployed, AI feature operating on real data |
| Enablement | 1 week | Training, documentation, handover |
Team: A compact team from our builders, scaled to SMB scope. Short delivery lines without enterprise discovery.
Approach
- Stack: typically Next.js or similar, Postgres, one AI provider behind a thin abstraction.
- AI: purpose-built per use case. Often RAG over the client's own data. Inference on EU infra by default.
- Deployment on the client's own domain. Hosting on EU infrastructure.
- Client owns code and data. No lock-in.
Proof
TODO(sanne): add an SMB custom-software case with a concrete outcome.
What this service is not
- An enterprise-integration project.
- A standard SaaS with a chatbot bolted on.
- A multi-month discovery without a working product.