Made for Swiss self-employed & SMEs
A notary in Zurich receives a 38-page purchase agreement on a Tuesday afternoon. A fiduciary in Geneva gets a 22-page tax ruling she has to act on by Friday. A doctor in Lausanne is sent a new insurance contract he never quite finds time to read. In every case, the document matters, the deadline is real, and the afternoon is already full.
Every notary, fiduciary, doctor and small office runs on documents it did not write: contracts, general terms and conditions, insurance policies, authority letters, bank documents, tax rulings. The people who draft them are paid to be thorough, so the documents are long, dense and written in a language that takes an hour to decode — an hour nobody has.
The result is familiar: we skim, we guess, and sometimes we miss the one clause that mattered. A calm AI routine can change this without changing how you work. The AI reads the document first and hands you a short, plain-language summary; you read that, then go back to the original only where it counts.
Why long documents get skimmed
It is rarely laziness. It is a mismatch between how documents are written and how busy people read:
- Density: legal and financial language packs meaning into long sentences and nested clauses.
- Volume: a single afternoon can bring several contracts, letters and policies at once.
- Mixed importance: ninety percent is standard boilerplate, but ten percent decides the whole matter — and it is never clearly marked.
The cost of skimming shows up later: a missed deadline, an overlooked liability clause, a fee you never noticed, or a client who asks a question you cannot answer because you never really read the terms yourself.
A calm routine in three passes
You do not need to read every page with a highlighter. Three quiet passes, with the AI doing the heavy reading, are enough.
Pass 1 — The plain-language summary
Paste the document and ask for a summary in everyday language. The AI extracts the structure, the parties, the subject and the main points, and writes them so a non-expert can follow.
Here is a contract I received. Summarise it in plain language, in about half a page: who the parties are, what is being agreed, what each side gives and receives, and how and when it ends. Write short sentences and avoid legal jargon. If you are unsure what a clause means, say so instead of guessing.
Pass 2 — Obligations, dates and open questions
Next, the AI pulls out everything that requires action from you: deadlines, notice periods, payments, liabilities, and anything unusual that deviates from a standard template.
Go through the same document again and list everything that requires an action from my side: deadlines, notice periods, payments, obligations, and any clause that seems unusual or stricter than a standard contract. For each point, give the date or condition, and mark with "to check" anything that is unclear or missing. End with a short list of open questions I should clarify before signing.
Pass 3 — The one-page client brief
When the document concerns a client, ask the AI to explain it in the client's language. A dense contract becomes a short, friendly note the client can actually understand — what was agreed, what it costs, what happens next, and what the client should do. That clarity is often half of good client service.
What AI does well — and what stays with you
- Reading the boilerplate: the AI works through the standard clauses so you only reread the parts that matter.
- Flagging the unusual: deviations, strict clauses and missing details stand out against the rest of the text.
- Translating into plain language: the same facts, written for a non-expert, without changing the meaning.
What stays with you is the judgement. The AI never signs anything and never advises a client on your behalf. It prepares the summary and the questions; you decide what the document really means and what to do about it.
A note on data and security
Contracts and official letters are confidential — they carry names, amounts, dates and sometimes health or financial details. They have no place in a public consumer chatbot. Use a local model or a zero-retention enterprise service, or remove the identifying details first and add them back yourself at the end.
The outcome
An afternoon of dense reading becomes twenty minutes of reviewing a clear summary. You catch the deadline before it passes, you answer the client's question on the spot, and you stop discovering — too late — that you had skimmed past the one clause that mattered.
