Not connected
A chat window next to the real systems does not change how work gets done. The value appears once AI can read and write where the work happens.
AI agency
An AI agency for SMEs: AI that fits into your systems, not next to them.
AI features, agents and automation, built with real software engineering instead of assembled from a no-code kit. Connected to your data and tools, with people reviewing wherever it matters.
- AI in products and platforms
- Agents and automation
- Human review built in
The starting point
Many AI projects never get past the demo.
Not reliable
Without clear scope, sources and review points, results vary and nobody trusts them in daily operation.
Not clear what is allowed
Data protection and the EU AI Act raise questions that stall projects before they start.
Services
Three ways EDS Labs brings AI into your business.
The core is building AI into products and platforms. Automation covers the recurring work around them. Orientation is the entry point when it is not yet clear where to start.
AI in products and platforms
AI features inside web platforms, customer portals, dashboards and apps, connected to your data and permissions.
- Assistants and drafting features
- Knowledge search over your documents (RAG)
- MCP connections to internal tools
- Agents inside web and app products
Automation and AI agents
Agents that take over repetitive steps in everyday workflows and hand decisions to people where needed.
- Email triage and reply drafts
- Document intake and extraction
- CRM updates and follow-ups
- Internal process automation
Entry and orientation
A structured look at where AI helps, what it takes and what to watch out for, before money goes into development.
- AI potential analysis
- Workshop with your team
- Technical assessment for the AI Act
Typical use cases
Where AI usually saves real time.
Examples of what AI projects for SMEs typically look like. Each project starts from your own workflows and data.
Document intake
Invoices, orders or forms are read, structured and passed on, with uncertain cases flagged for a person.
Internal knowledge assistant
Answers from manuals, policies and project documents, with sources, instead of searching shared drives.
Email triage
Incoming emails are sorted, summarised and prepared as reply drafts that a person sends.
Support assistant in the portal
Customers get answers inside the customer portal, with a clean handover to the team for anything complex.
CRM agent
Meeting notes and emails turn into CRM entries, tasks and follow-ups without manual copying.
Internal tools via MCP
AI gets controlled access to internal systems through Model Context Protocol, with permissions and audit logs.
How we start
Three steps from first idea to running system.
- 01
Free intro call
We talk about your workflows, systems and goals, and whether AI is a sensible fit.
Free, 30 minutes
- 02
AI potential analysis
A structured analysis of your processes and data, with a written assessment and a concrete roadmap for a first pilot.
Fixed price on request
- 03
Pilot, then operation
One focused use case built with real data. Once it holds up, it moves into daily operation and is extended step by step.
Scoped per project
Why EDS Labs
Proven in production, not on slides.
EDS Labs is a focused studio. What counts is what runs in production.
AI in our own products
uLiquid Desk, a live EDS Labs product, uses AI-assisted market analysis with paper mode and explicit confirmations, so users stay in control.
Visit uLiquid DeskAgents with human review
EDS Labs works with AI coding agents every day. Architecture, product decisions and code quality stay under deliberate human review, and the same principle applies to client AI systems.
Web, app and backend from one source
AI rarely stands alone. The surrounding product, APIs, portals and apps are built by the same studio, without handovers between vendors.
A focused studio
Direct contact with the person building your system, lean communication and clear decisions, supported by a freelancer network where a project needs it.
Privacy and AI Act
Data protection and regulation are planned in, not bolted on.
- Model and hosting are chosen per project to fit your data and requirements.
- Only the data a use case actually needs is processed.
- Review points, logging and clear responsibilities are part of the design.
- AI Act requirements are assessed technically and organisationally for each use case.
The AI Act assessment is technical and organisational guidance, not legal advice.
FAQ
Common questions about AI projects.
What does it cost to get started?
The intro call is free. The AI potential analysis is offered at a fixed price that depends on scope and is quoted on request. Pilots and implementation are scoped per project.
Do we need our own data?
Not necessarily. Many use cases work with existing documents, emails or system data. The potential analysis clarifies which data a use case needs and in what quality.
Which AI models are used?
That depends on the use case, the data and the requirements. Options range from models with data processing in the EU to self-hosted open-source models and established providers. The choice is made per project.
How long does a pilot take?
A focused pilot usually takes a few weeks, depending on the systems involved and the available data. The exact timeline is part of the potential analysis.
Does EDS Labs work remotely with companies in Germany, Austria and Switzerland?
Yes. EDS Labs works remotely with companies across the DACH region and internationally, in German or English.
What happens after launch?
AI systems need monitoring and adjustments as data and requirements change. EDS Labs can continue to operate, improve and extend the system after launch.
Find out where AI helps in your business.
Book a free 30-minute intro call. Describe your workflows and systems, and you get an honest first assessment of where AI makes sense and where it does not.