AI for banks.
Sovereign, auditable, FINMA-ready.
Use AI without leaving supervisory questions open: AIgent processes banking data in Switzerland or on-premise, logs every agent step, and lets you choose the model by data class.
What does AI give a regulated bank?
For banks, AI resolves a concrete tension: high productivity pressure on document-heavy work, paired with a regulatory frame that demands traceability and data control. AIgent addresses exactly that, recording every agent step with model, time, and cost in the audit trail, while highly sensitive data uses an in-house model.
Banking secrecy under Art. 47 of the Swiss Banking Act does not end at the cloud border. That is why AIgent runs in Swiss data centers under Swiss law, or fully on-premise, independent of US hyperscalers and outside the US CLOUD Act.
What supervisors expect from AI.
With supervisory guidance 08/2024, FINMA set out expectations for AI governance: accountability, data quality, traceability, and control. AIgent supports this with transparent agent reasoning instead of a black box, a complete audit trail, and central management of rights and models.
Add the outsourcing requirements and the DORA relevance for operational resilience. The on-premise option and per-data-class model choice give you the control an outsourcing review expects.
What an agent handles in the bank.
Email triage
Prioritise incoming requests, route them, and prepare them with context.
ViewPrepare KYC documents
Pre-fill forms and onboarding documents from existing data.
ViewPre-analyse cases
Enrich items with reference cases before they reach the team.
ViewReports in corporate design
Generate reporting drafts and documents in your corporate identity.
View- Payment terms compliant
- Collateral complete
- !Clause §7 flagged
Common questions: AI in the bank.
- Does AIgent meet FINMA outsourcing requirements?
- AIgent is built to support an outsourcing review: you keep control over data location, access rights, and model choice, and every processing step is documented in the audit trail. For the highest level of control you run AIgent on-premise on your own infrastructure, which removes classic outsourcing to a third party. The supervisory assessment in each case remains your compliance team’s task; we provide the technical preconditions.
- Is on-premise operation possible?
- Yes. AIgent installs via a Helm chart on your own Kubernetes cluster, up to air-gapped with no internet connection. It is the same product as the Swiss cloud, just entirely in your environment. Neither data nor models leave your infrastructure, and you connect your directory with single sign-on and role synchronisation.
- Who can see our data?
- Only you and the people you authorise. AIgent does not use your content for model training and produces no analytics from it for third parties. You manage access and roles centrally, and every access is traceable in the audit trail.
- Which models run in Switzerland?
- You choose the model per agent by data class. Highly sensitive data such as banking information is processed by an in-house model or in the Swiss cloud, while non-critical, non-personal tasks may run on powerful international models. The model follows the sensitivity of the data, not the other way around.
Sovereign AI. For every discerning company.
On your use-case, in your environment, within a clearly defined scope. See measurable results within a few weeks.
Secure early access. No per-seat fees. No lock-in.