Services
System, take over from here
Four core competencies, one solid technical base. We automate your routine processes, make internal knowledge instantly available, build high-performance web platforms and improve your visibility in modern AI search engines.
Our focus areas
Corporate memory
AI knowledge systems
Hallucination-free answers based on your own documents.
An internal knowledge system on a RAG foundation (Retrieval-Augmented Generation): it answers questions about offers, contracts and handbooks reliably and with exact source citation. Depending on the data structure we use vector search, relational databases or knowledge graphs. We measure the hit rate precisely, in black and white.
Via multimodal RAG (Visual RAG) the system also captures tables, charts and plans, not just plain text.
Learn moreAutomation
Automated workflows
Routine tasks run flawlessly in the background.
Smart n8n workflows and integrated language models (Anthropic API) that relieve your team. The systems screen incoming requests, pre-qualify leads, prepare documents and connect your existing software, fully automated and reliable.
Learn moreNext.js / Headless
Modern web platforms
Fast, low-maintenance and absolutely secure.
Websites built on Next.js and a headless CMS (Strapi), server-rendered in milliseconds. 100% EU-hosted, privacy-compliant without third-party tracking (Plausible) and structurally free of the typical security risks of classic WordPress installations.
Learn moreGEO / AEO
AI search optimisation
So algorithms recommend you.
Generative Engine Optimization (GEO/AEO): we structure your content so that ChatGPT, Perplexity and Google AI surface you as a trustworthy source. With clean semantic data structure (Schema.org) and continuous monitoring we secure your visibility. Ask our chat in the corner, it works exactly like this.
We know exactly how AI search engines select data, because we build such RAG systems ourselves, as developers.
Learn moreTechnical depth
What sits behind our projects.
Six technical areas that keep coming up in our work — explained briefly.
- Search and knowledge systems
We combine full-text search with semantic vector search and merge both result lists using rank fusion. The hits are then re-sorted (reranking). Depending on the query and the data, a router decides whether vector search, an exact database query, a knowledge graph or a cache fits better. This principle runs in production every day in jukeep.
More info- Automation in day-to-day operation
Multi-step n8n workflows with monitoring and automatic retries on errors, connected to CMS, object storage and messengers. A typical sequence: a message or document comes in, the data is captured in a structured form, a draft is created and is then approved by a person. At WOOD CUBE this pattern has been part of daily business since 2021.
More info- Websites and platforms
We develop with Next.js, server rendering and a headless CMS, with our own deployment pipeline and hosting in the EU. Before a relaunch we review the technology, performance and findability in classic search and in AI answers (SEO/GEO). The new trainCo website came about this way, among others.
- Custom applications instead of off-the-shelf software
When standard software does not map a business process cleanly, we build the application that fits, with its own data model and its own interface. Examples are the geo application for GESIBA and the yoorda QR ordering and POS system.
- Understanding documents and processing them further
We combine text recognition (OCR) with language models in order to read content out of documents and process it further in a structured way. At yoorda this runs in production when menus are read in and translated. For scans, tables or plans without a reliable text layer we use Visual RAG and evaluate the page directly as an image.
More info- Operations and infrastructure
EU hosting, server migrations, deployment pipelines, monitoring and backups are part of day-to-day operation for us. After launch the systems stay in our care — with some clients for many years.
Our approach
Our principles.
We choose the tech to fit your data.
How your data is structured determines which solution makes sense. Sometimes that's a large language model, often something leaner is enough.
Measurable results over gut feeling.
We measure how accurately and stably the system answers, and document it. That way you see what you can rely on.
Sometimes you don't need AI at all.
Where a classic database query is faster, safer and cheaper than a language model, we tell you so before the project starts.
Built for production.
We build on a maintainable, documented stack (Next.js, PostgreSQL) and keep the systems running in everyday operation.
Expertise
Specific components that integrate seamlessly into your IT.
Voice AI
Phone assistants can now take calls, record what a caller needs and enter appointments, for example. If that makes sense for your business, we could build such a solution around your specific workflow, using speech recognition and language models.
MCP servers
Secure connection of internal data sources to AI assistants via the Model Context Protocol (MCP). This website runs its own MCP server.
Custom software
APIs and interfaces that connect existing systems cleanly with each other — even when there is no ready-made module for it.
AI support
Continuous system care and optimisation through an embedded AI engineer on call.
Tailored to your industry.
Proven technology that is adapted to your operational requirements.
Law & firms
Files, deadlines and precedents findable again, kept in the EU.
SMEs in general
Scattered company knowledge searchable in seconds, with sources.
Construction & modular
Quotes, calculations and project documents close at hand.
Trades & crafts
Faster quotes, scheduling and order history on request.
Housing
Custom tools for recurring administrative workflows.
Education & training
Course knowledge made searchable, administration simplified.
A bit about us
A few key facts.
No hype, just a few facts about us:
20+
years web & e-commerce
2001
webse brand since
4
own AI products
100%
EU-hosted & GDPR-compliant
We build technical standards in from the start:
No more manual reporting. The system delivers automated site reports and visualisation in seconds.
Mag. Mag. Roland Böhm, LL.M.
Member of the Management Board, gesiba.at
FAQ
What companies ask us about our services
- Where should we start?
- With the area that costs the most time or nerves today, often a knowledge system (finding things) or automation (taking over routine). We sort that out together in a conversation.
- Do we have to pick one service?
- No. The areas work together, for example a knowledge system with automation. We start where the benefit becomes visible fastest.
- Do you work EU-compliant and at a fixed price?
- Yes. We host in the EU and agree clear fixed prices per step, so you know the cost upfront.
- Who looks after operation and further development after launch?
- We do. After go-live we take on maintenance, monitoring and further development. We check whether answers and workflows still fit in daily use, respond to changes and look after servers, updates and backups. Some clients we have supported this way for many years.
