# AI Deployment From pilot to an application that runs every day. We build and run AI applications for foundations, public administrations and universities. ### Agents and automations The AI takes over recurring steps: summarising documents, checking entries for completeness, moving information between systems. You decide what runs automatically and what a person decides. ### Integration and administration We connect the AI to your data, specialist applications and existing interfaces. Where no suitable connection exists yet, we build the integrations or MCP servers it needs. ### Governance and operations Access rights, logging and quality controls, so it stays traceable at all times who triggered what, and on what basis. We solve concrete problems, test the benefit in practice and move into operation what has proven itself. ### Workshop and pilot Together we look at your data and constraints and find the use case that brings the most and carries the least risk. Out of that comes a working pilot, including interfaces and access management. ### Roll-out and expansion Once the pilot is validated, we bring it into your team's daily work, improve it on the first experiences and take on the next use case. ### Operation and monitoring When everything runs stably, we take over operation and monitoring. You keep control of data and permissions. ### Data protection Your AI application runs entirely in Switzerland. Sensitive data stays under your control and is never used to train models. ### Security and control Every automation gets only the access its task requires. Logging and monitoring make every action traceable. ### Ease of use We reduce technical complexity. Our experience in UX and product development makes sure the solution is used every working day. ### What is an MCP server? MCP (Model Context Protocol) is an open standard for connecting AI applications. An MCP server provides an AI agent with the tools and data it needs. That lets it search documents, retrieve information from a specialist application or trigger a process, for example. ### What are AI agents? AI agents support you with tasks that consist of several steps. They access defined data and applications and carry out the individual steps on their own. You might ask: “Show me every application from the last ten days.” The agent finds the matching applications in the system and lays the results out clearly. Along the way it can combine information from different systems and process it further. ### Is AI deployment part of the Fundament app or a separate service? You can book AI deployment independently of the Fundament app. If you already use Fundament, we build straight on top of it. ### What can AI agents be used for? AI agents suit tasks where information has to be found, checked, brought together or processed further. They can prepare documents from existing data, summarise applications and dossiers, or collect information out of several specialist applications. Recurring steps can be simplified the same way. Which tasks are a fit for your organisation is something we work out together, from your existing processes. ## Where would AI make a difference for you? In a first conversation we look at your processes and draw up a non-binding list of measures for you. [Request a meeting](/en/contact) - [Alle Welt redet über KI. Was hält unser Tech-Lead davon?](/en/journal/alle-welt-redet-ueber-ki-was-haelt-unser-tech-lead-davon.md): (Viel.) - [Was KI im Förderwesen wirklich verändert](/en/journal/was-ki-im-förderwesen-wirklich-verändert.md): Künstliche Intelligenz ist bei den Stiftungen angekommen. Wo bringt sie echten Mehrwert - und wo nicht? Wir hätten da ein paar Gedanken.