"So I made a profit and there's nothing in my account?" AI can help you answer without jargon. What it can see of your client's accounts, and what you are better off keeping to yourself.
"So I made forty thousand euros in profit, and there's nothing in my account?" The question comes up at every year-end accounts meeting, often from a tradesman looking at his first balance sheet. You know the answer: the stock, the customers who haven't settled their invoices yet, the loan repayments. The hard part is putting it without jargon, in ten minutes.
AI is very good at finding those words. The question is what you can show it of a client's accounts.
Not to just any AI. In many countries, accountants are bound by professional secrecy, and breaching it can be a criminal offence. In Switzerland, auditors are covered by Article 321 of the Swiss Criminal Code, and the fiduciary who keeps the books owes the client discretion and compliance with the FADP; in the European Union, the GDPR applies as soon as the figures relate to an identifiable person. Pasting a named balance sheet into a consumer AI tool means passing it to a third party that may keep it, sometimes under foreign law that gives its authorities access. Check your firm's policy and your engagement letter too.
Rarely. A turnover, a sector and a town are often enough to recognise a small business, and a director's loan account or a manager's salary points to a person. Removing the name is pseudonymisation, not anonymisation: our guide on anonymisation and pseudonymisation explains the difference. For teaching purposes, round the amounts and keep only the items you need, or stay on a confidential model.
In IA Confidential, questions are handled by default by confidential models: open models installed on servers in Switzerland, which pass nothing on to third parties. To put a balance sheet into plain words, they are all you need. If you choose an external model, the Confidentiality filter first detects sensitive data and offers to answer with a confidential model, message unchanged, or to send an anonymised version that you review before it goes.
Describe it once. Create a "Client meetings" space and set the rule in its instructions: no technical term without a definition, everyday comparisons, one page at most. These instructions take precedence over your general preferences. Then start simple:
Explain to the manager of a bakery, without jargon and in five sentences, why a business can show a profit and still have almost no cash. Use the example of stock, customers who settle at sixty days and the repayment of a loan.
Yes, and confidently: a ratio miscalculated, a total copied wrongly, profit confused with cash, a plausible but wrong explanation of a change. Ask it to cite every amount it uses and to separate what it reads from what it infers, then check everything against your accounts. What you tell the client is still your advice, and your responsibility.
Here are the main headings of the balance sheet of a joinery business with six employees, in thousands of euros, over two financial years. Explain the three most significant changes to a director who is not a finance person. For each one, cite the amounts you use and distinguish what the figures show from what you are assuming.
Yes. The file is read in your browser, and nothing is stored on our servers. If the question goes to an external model, only pseudonymised extracts are sent; if the document is judged too identifying, the assistant tells you and offers to have it handled by a confidential model, or to send only the question. It can then draft the meeting note as a document: editable, printable as a PDF, exportable to Word or OpenDocument with a fresh layout.
Using the attached annual accounts, write a two-page note for the director: what the balance sheet says in three sentences, a table of the five items that changed most with a simple explanation, two points to watch in the coming year and three questions to ask at the meeting. Do not invent any figure that is not in the accounts, and flag the ones you are missing.
With a confidential model, the balance sheet, your questions and the answers are processed on servers located in Switzerland, and nothing is kept on our side: the history stays in your browser.
With an external model, what is sent depends on your decision, once the filter has run: an anonymised version, pseudonymised extracts, or the message as it is if you accept that. In the Privacy panel, you can also specify what must always be masked, such as your clients' names.
On a shared computer at the firm, the history can be read by anyone using the same browser session: keep a personal session and close it when you leave. And the client's file stays in your accounting software, not in the chat.
Before pasting a trial balance or a general ledger, reread the list of data you should never paste into an AI and our habits for using AI at work safely.
To try it on a fictitious balance sheet, with nothing to install: open the chat.
