AI integration for companies
Most AI projects fail not because of the model but because nobody defined which work it is supposed to replace. We start with the workflow, not the model: only once it is clear which hours or which error rate should disappear do we pick the technique that delivers it — sometimes a language model, sometimes twenty lines of code are enough.
Who this is for
- Companies with many internal documents where staff search for answers every day.
- Businesses manually processing incoming documents such as invoices, enquiries or forms.
- Teams answering the same customer questions over and over.
- Organisations that want to use AI but must demonstrate data protection and traceability.
Prices and timelines
Fixed price ranges, no open-ended hourly billing. The final fixed price is in the quote before any work starts.
| Service | Price range | Duration |
|---|---|---|
| RAG system — answers based on your own documents | 5 500–18 000 € | 4–9 weeks |
| AI agent — multi-step workflows with tool access | 9 500–35 000 € | 6–14 weeks |
Prices excl. VAT — small-business exemption under § 6 (1) 27 Austrian VAT Act. You can get a first estimate in two minutes with the calculator on the homepage.
How we work
Benefit check
We look for the one workflow where AI measurably changes something — and say so when we cannot find it.
Proof of concept
A small test on your real data before you invest in a full implementation.
Integration
Built into the systems your team already uses. Not another tool nobody opens.
Measurement
A before-and-after comparison on a number agreed before the project started.
Technology
Models
Chosen per task, data-protection requirement and cost — no single model for everything.
Retrieval (RAG)
Vector database with source attribution: every answer points to the document it came from.
Data protection
European processing where the GDPR requires it. Local models when data must not leave the building.
Traceability
Logging of input, source and output — so a result stays verifiable.
From our projects
BookForge — AI book creator
Trilingual tool to draft AI-assisted books: outline, chapters and editing in one flow.
Digital wine list
Digital wine list webapp for GansFein. Filter by grape, pairing and region, built on the brand identity we also designed for them.
Frequently asked questions
What does an AI project cost?
A RAG system on your own documents runs between 5,500 and 18,000 euros, an agent with multi-step workflows between 9,500 and 35,000. On top come ongoing model costs, which depend on usage volume and which we project in advance.
Does our data stay in Europe?
If that is required, yes. We choose the processing according to your data-protection situation: European providers, or a locally run model when data must not leave the building. That is settled before the project starts, not after.
How do you stop the AI from making things up?
Through retrieval rather than free generation: the system answers from your documents and states the source. Where no source exists, there is no answer. That is less impressive than a model that always says something, but in production it is what counts.
What if AI is not the answer to our problem?
Then we say so in the first conversation. Some of the enquiries we get are better solved with a database query or a clear form than with a language model. That also costs you correspondingly less.
How long does implementation take?
A RAG system four to nine weeks, an agent six to fourteen. Before that comes a short proof of concept on your real data, so the investment decision is not based on a demo with sample data.
Let us talk about your project
A conversation where we understand the problem and tell you honestly whether and how we can help. Free, no obligation, in German, English or Spanish.