The senior layer of your firm is doing junior work. We install the agents that take it on, so your senior people spend their time on the calls only they can make.
Your data stays in Switzerland or the EU.
We build the secure data infrastructure first. Hosted in Switzerland or the EU, governed by the Swiss Federal Act on Data Protection or the European GDPR. Every agent runs on a foundation that was already built to your jurisdiction.
AI installed inside your firm.
We don't replace your systems. We add agents on top of the stack you already run, so they work on your data, next to your people, inside the tools they already use.
The execution off their desk. 2–4× output.
The agent works through the whole process: it pulls the data, drafts the memo, reconciles the numbers, routes it for sign-off. Your people decide and validate. They stop being the one typing.
The office keeps a memory, your people can ask it anything, and agents run on a schedule. Each one reads from the same foundation.
Every letter, IC pack and decision becomes a structured record. The agents read from it and write back to it, so the context compounds instead of getting lost in someone's inbox.
Your people ask a question in plain language and get an answer grounded in your own data and memory, with the sources attached. Nothing leaves your systems.
You set a task once and it runs on its own, overnight or weekly, on the schedule you decide, and it leaves an audit trail. This is the proactive layer a chat box never reaches.
Build the foundation once. Reporting, dashboards, and workflows all read from the same memory the platform has been building. Whatever your firm adds in three or five years, each new initiative reads from the same foundation. No re-ingestion. No re-migration. No vendor lock-in at the foundation layer.
Hosted in Switzerland or the EU, under your own keys.
Never tied to a single US AI lab — you keep the choice of model.
Whether it's empowering every employee to build the dashboard their department actually needs, or agents that finish a full task end to end. Two patterns we've already installed. The rest gets shaped to your firm in Discovery.
The agent maps the data sources, infers the metrics that matter, drafts the dashboard, lets the employee tune it. No central BI bottleneck. Each person, the view they actually need.
The agent drafts posts from your decision memory, queues them on LinkedIn and your other channels, tracks engagement, learns what lands. The team approves, doesn't write.
Consumer AI is everywhere. Enterprise AI gates it. We build the deployment that fits the firm.
Eight to twelve weeks. We don't rip and replace. Your stack stays. The platform stays with you.
step-1-knowledge-graph.svgWe map your firm. Who hands what to whom, where the friction lives, which artefacts get re-touched ten times. The output is a knowledge graph of how the work actually runs, plus a shortlist of processes worth handing off.
Whether your team is fifteen or two hundred, the bottleneck is the same. A senior layer doing judgement work, surrounded by execution that consumes them. Memos. Reporting. Sourcing. Drafting. We take it off their desk.
Deal screening. IC memos. Portfolio reporting. Off the analyst’s desk.
First drafts. Precedent search. Document review. Off the associate’s desk.
Underwriting. Deal flow. Asset reporting. Off the operator’s desk.
A real estate group operating across seven countries, with growth plans that have outrun the shared picture of how the company actually runs. In Phase 1 we mapped the operational reality end-to-end, across every function, surfacing where work breaks, duplicates or sits with a single person. Phase 2 turns that map into a cleaner data layer and the first pilot agents.
Discover caseAn industrial operator with thirty years of process knowledge spread across binders, drives and people about to retire. We built the knowledge layer that captures every decision as a structured record and surfaces it where the next decision is being made. Institutional memory, finally readable, finally inherited.
Discover caseYour data lives in the EU, in Switzerland, or on your own servers — depending on your setup and your current needs. Stored in jurisdiction-specific infrastructure, governed by the Swiss FADP or the European GDPR. It does not leave.
A Swiss or EU court order would be required. The US CLOUD Act does not reach Swiss-owned infrastructure. On the dedicated tier, the decryption keys are held by you, so even a court order to us would not produce readable data.
No. Your data runs your agents. It never feeds training sets, ours or anyone else's.
Eight to twelve weeks from discovery to running in production. The platform stays with you afterwards.
Everything exports. Documents, decisions, memory, audit logs, as ordinary files in standard formats. No re-licensing. No vendor handover process.
Thirty minutes. We show you an agent running on your data. Not a deck, not a roadmap. If it fits, we scope. If not, we tell you straight.
Bring one workflow that eats your senior team's week. We show you what comes off their desk, on your data, on Swiss or EU infrastructure.
Book a callIn a world moving this fast, the work needs curiosity wide enough to absorb what's changing and depth enough to solve what isn't. Technical fluency, complex problem solving, real industry time. The four of us share all three, and all four of us studied at ETH.
CS at ETH. A year as technical lead at an AI startup.
Mechanical Engineering at ETH. Founder of Alpine Valley and Vita Service.
CS at ETH. Six months at a software consultancy.
Physics at ETH. Two years in the financial industry, the last three months as AI lead.