Software should last.
Alpid is an independent technical practice in Switzerland. We take over the systems your business runs on, we build AI agents for the routine work, and we test agents before they are trusted with real work.
Lasting is the hard part. Software should outlive the team that built it, survive its next change of platform, absorb years of new requirements, and still be worth improving. That is the standard we build to, and we keep what we build up to date for as long as it runs.
How we build has changed. We use today’s AI technology, agents included, and it carries the routine work well. The responsibility does not move: it stays with real people, seasoned engineers, Swiss-trained. An engineer answers for every change that ships.
We come to this from software engineering and from agentic AI. Careers spent on software where mistakes cost money: market infrastructure, heavy industry, commodity trading, and multi-agent AI systems in production. Alpid was founded by Niels Escarfail.
That background is why we offer three practices. Modernization: we rebuild the system nobody dares touch and ship improvements weekly; what we build is yours, including the IP. Agents: for the routine work, we build one to carry it, under permissions decided first, with every action logged. Audit: before an AI agent goes live, we test it, attack it, and attest what broke and what held.
The standards we answer to.
Our audits run against external standards: AIUC-1, ISO/IEC 42001, EU AI Act readiness. Public yardsticks, set by others, so the result means something outside our walls.
And if you are in regulated finance, you have seen this kind of work before. Threat-led testing, the TIBER-EU and DORA TLPT exercises regulators ask for by name: our audits follow that tradition, extended to AI agents, so the report fits a process your risk team already runs.
In Switzerland the expectation is already written down. FINMA Guidance 08/2024 records that only a few supervised institutions carry out an independent review of the entire model development process by qualified personnel. Our audits answer that expectation for AI agents.