We're building the AI co-worker that runs Europe's most complex construction projects — transparent, auditable, end-to-end.
Live in production on a major autobahn program and an S-Bahn transit program. Pre-seed closed. Scaling the product to our first twenty customers right now.
The construction industry loses $1.6 trillion annually to inefficiencies. Not because people are careless — but because critical knowledge is trapped in PDFs, meeting transcripts, and inboxes. Every new project starts from scratch. Every decision is remade from incomplete context.
We're building the opposite: project management software where AI runs as a co-worker — executing workflows end-to-end, surfacing risks, and compounding knowledge across every project we touch. The moat isn't the model. It's the project memory no competitor can replicate.
75% of the infrastructure that will exist in 2050 doesn't exist today. We're going to build it better.
Three core problems. All in production. None of them have Stack Overflow answers.
One summary doesn't fit all. Structural risk in an RFI means something different than in a cost review, which means something different in a schedule reconciliation. We build multi-agent harnesses — specialized extraction, reasoning, and evaluation stages — that route 400-page tender documents and protocol archives through the right pipeline with the right context.
If you've read Anthropic's or LangChain's writing on agent harnesses and thought yes, that's the hard part of shipping production agents — this is where you'd live.
We started with meeting transcripts. We're building a decision graph that grows with every project — tracking not just what was decided, but why, by whom, against which alternatives, and with what outcome. That graph feeds the next project. Every closed workflow makes the next one faster.
This is the operational-continuity layer no ConTech player is building. The hard part isn't retrieval — it's deciding what signal to keep.
What should the system remember? Forget? Surface at which decision point? There's no clean top-k answer when a project spans 10 years and touches 50 stakeholders. Our agent already completes entire workflows autonomously end-to-end — and the data architecture has to survive five generations of AI capability evolution without rewrites.
This is an open research problem we're solving in production.