Made for Swiss self-employed & SMEs
Most rejected offers are not badly written — they are unclear. The client cannot see exactly what they will get, what it will cost, or why the price is fair. AI can draft and structure an offer in minutes, but the offer closes because the scope, the price and the trust are right — not because the prose is polished.
If you write offers regularly, the pattern is familiar: you spend an evening on a quote, send it, and hear nothing back. The problem is rarely your wording. It is what the offer leaves out — and that is exactly where a careful AI workflow helps.
Why careful offers still get turned down
When a client goes quiet after receiving an offer, it is usually for one of four reasons:
- Fuzzy scope: the client cannot tell what is included — and what would cost extra.
- An unexplained price: a total without a breakdown reads as arbitrary, even when it is fair.
- Unanswered questions: the offer answers the request you remembered, not the one the client actually wrote.
- Generic text: visibly recycled paragraphs signal that the offer was made in a batch.
Notice that none of these is a writing problem. They are clarity problems — and clarity can be built into the drafting process.
What AI does well here
A language model is genuinely useful for the mechanical half of an offer:
- Restating the request. Paste the client’s email or chat messages and have the AI summarise what they asked for, in their own words. An offer that visibly answers the client’s actual request starts ahead of most competitors.
- Keeping the structure consistent. A fixed skeleton — task, approach, deliverables, price, next step, validity date — means nothing important gets forgotten when you are tired.
- Drafting option tiers. A small, a standard and a comprehensive variant helps the client choose instead of stalling. The AI writes the three descriptions; you set the three prices.
- Explaining the price in plain language. What the client buys, what each part costs, what drives the total up or down.
What AI cannot decide
Keep three things firmly human. First, pricing: a model knows nothing about your margins, your capacity or the going rates in your region. Second, scope promises: models tend to add impressive-sounding deliverables you never intended to include. Third, commitments: deadlines, warranty terms and cancellation rules must come from you, not from a draft.
A practical offer workflow
- Summarise the request. Paste the client’s messages and ask for a short summary plus a list of open questions.
- Answer the open questions. Fill in what you know — and ask the client what you do not. An offer built on guesses is a discount in disguise.
- Draft with your own material. Give the AI your structure, an anonymised past offer and the confirmed facts. Use placeholders for prices and deadlines.
- Check the substance. Every deliverable priced, nothing promised that you cannot deliver, every figure from your notes — not from the model.
- Make it personal. Add the one detail that shows why you are the right fit, and close with a concrete next step and a date.
“Draft this offer from the client’s request below and my example offer. Use placeholders for prices and deadlines. Do not add deliverables I have not listed. Flag anything the request leaves unclear.”
What to measure
Track three numbers: offers sent per month, reply rate, and time spent per offer. If a 30-minute offer replaces a two-hour one, you will send more of them — and following up stops feeling like a chore. That is usually where the real gains appear, long before any wording magic.
And when an offer is accepted, the next step should be just as clean: send a proper Swiss invoice with a QR-bill, correct payment terms and clean numbering, so the momentum is not lost in admin.
