Our Focus
The three areas where we do our deepest work, from first prototype to production.
AI Agent Automation
Automate complex, multi-step workflows with AI agents that coordinate tools, make decisions, and hand off to a person when it matters.
Guided Intake Solutions
Conversational intake that guides customers to the right details up front, resolves simple requests on the spot, and turns the rest into clean, structured cases.
Document Processing Pipelines
Pipelines that read, classify, and validate incoming documents, then feed structured data straight into your downstream systems.
Selected Work
Systems we have built, and are building, for insurers and the public sector in the DACH region.
AI-Guided Claims Intake
A guided intake that walks claimants through their case, determines which documents are needed, reads invoices and files as they are uploaded, and requests anything missing directly in the conversation.
Automated AI Brand & Compliance Checks
A dashboard that checks over 126 criteria across thousands of websites and social media profiles, from LinkedIn and Instagram to Facebook and Google Business.
Guided Intake on Local LLMs
A guided intake for the public sector, built on our guided intake solution and powered by locally hosted LLMs, so sensitive data never leaves the organization's own infrastructure.
What We Build
Data Intake & Processing Pipelines
Structuring, quality gates, plausibility checks, query logic, and handoff to downstream systems.
Agentic Workflows & Automation
Copilots and automations with guardrails, logging, approval processes, and traceable process logic.
Training & Cloud Deployment
Scalable training, data processing, MLOps, and production operations on AWS, Azure, and GCP.
From Idea to Production
In months, not years.
Online Meeting
45 min to 1 hourWe discuss your goals, challenges, and first ideas for using AI. Together, we identify where AI can create real value.
Scoping Workshop
2 to 3 hoursWe sharpen the use case, show relevant demos where helpful, and define what will be built, how the implementation will work, and what costs to expect.
Build Phase
3 to 6 monthsWe develop the solution step by step, show regular progress, and integrate it cleanly into your existing systems.
Meet the Team
Three ETH Zurich engineers bringing AI from the lab into production.
Andre Emmenegger
Agentic Systems & Platforms
- MSc CS, ETH Zurich
- Time Series & LLMs Research
- Ex-BSI Engineer
- Published at Disney Research
Darius Doongaji
Innovation & Intelligence
- MSc CS, ETH Zurich
- Full-stack Engineer at SF Startup
- Lesion Detection Research
- Built Robust AI at Medical Startup
“We connect current AI research with business practice and turn applied research into systems with measurable results.”
Common Questions
We implement least-privilege access, auditable data flows, and human-in-the-loop validation where it matters. All systems can be deployed on-premise or in your own cloud environment with full data sovereignty.
Discovery: 1-2 weeks. MVP development: 2-6 weeks with weekly demos. Full production rollout typically takes 8-12 weeks in total, depending on complexity and integration needs.
We work either project-based or on a retainer. After the scoping workshop, you'll have a clear roadmap with effort estimates. We focus on delivering value, not billable hours.
Yes. We build cloud-agnostic solutions that can run in AWS, Azure, GCP, or on-premise infrastructure. You maintain full control of your deployment environment.
Every project starts with defining measurable KPIs tied to business outcomes (e.g., time saved, error reduction, margin improvement). We track these metrics throughout development and production.