AI for fiduciary & finance
AI for fiduciary, accounting and controlling in Switzerland
In fiduciary and finance work, it is not the posting that takes time but everything around it: matching receipts, answering queries, finding anomalies before the closing, preparing figures for management. This is exactly where AI agents that access Abacus, Bexio or SAP directly, with the permissions of the respective person, take work off your hands.
What fiduciary firms and finance departments use AI for
- Receipts and account assignment
- Reading incoming receipts, proposing supplier and account, flagging inconsistencies. Posting happens after review by a human.
- Client and customer questions
- “Which invoices from client X are still open?” in plain language, directly from Abacus or Bexio, with client separation through the existing permissions.
- Preparing closings and VAT
- Compiling missing receipts, unusual postings and variances from the previous year before the closing work begins.
- Controlling and forecasting
- Explaining variances against budget, producing liquidity and revenue forecasts. Dr. Maher Hamid has worked with forecasting and machine learning for finance for years.
Confidentiality and the FADP
Client data is the capital of a fiduciary firm. That is why the rule is: the data stays in the source system, the MCP server reads only what is necessary per request, and each request runs with the permissions of the requesting person. Platform providers generally require a data processing agreement; for providers abroad, the rules on disclosure abroad also apply.
Where that is not sufficient, the models run locally or in a Swiss data centre. We provide the information your data protection assessment needs. This does not replace legal advice.
Your team carries it forward
We train clerical staff and management on your own workflows, not on textbook examples. Dr. Maher Hamid teaches AI and LLMs at the Haufe Akademie, the Klubschule Migros and the IBAW. After the project, your team knows the tools and can extend them themselves.
Limits we state up front
- No postings, payments or filings without review by a human.
- No circumvention of client separation in the source system.
- No payroll data unless you explicitly approve it.
- No tax or legal advice by the AI.
Frequently asked questions
Does this work with Abacus and Bexio?
Yes. Bexio offers a public REST API with per-person sign-in; Abacus provides interfaces depending on the version and licensed modules. Other systems such as SAP Business One or S/4HANA can also be connected.
Does every clerk see all clients?
No. Each request runs with the permissions of the requesting person in the source system. What someone is not allowed to see there, they do not see in the chat either.
Does client data stay in Switzerland?
This can be planned that way: the MCP server in your infrastructure or in a Swiss data centre, language models with data residency or running locally on your own hardware. Which option fits depends on your clients and your contracts.
What is the best way to start?
With a potential analysis: we talk to the people who do the work, measure where time is lost, and calculate the potential in hours and Swiss francs with the calculation method disclosed. This is followed by a pilot on a clearly defined workflow.
Let’s check whether it holds up for your systems.
The feasibility study takes one to two weeks and ends with a document you can keep working with even without us. If an interface cannot meet your requirement, it says so.
Request a first call