Made for Swiss self-employed & SMEs
A one-person business does not need a 60-page market report to make a better decision. It needs a few trustworthy signals, collected for a specific question. AI can help organise that evidence, but it cannot turn a handful of webpages into proof that customers will buy.
Start with a decision, not a prompt
“Research my market” is too broad to produce useful work. Pick a decision you actually face: which service to promote next quarter, whether a new client segment is worth pursuing, or how to describe your offer more clearly. Write down what would change your mind. This keeps the research small enough to finish and makes it harder for a fluent AI answer to sound like a conclusion.
For a Swiss freelancer, the question may also be local: does demand differ between German-speaking and French-speaking clients, between cantons, or between urban and rural areas? Treat those as questions to investigate, not assumptions to bake into the prompt.
Build a small evidence file
Gather a modest set of public, relevant sources rather than asking a model to invent a market overview. Depending on the decision, that might include your own anonymised inquiry themes, public competitor service pages and prices, official Swiss or cantonal statistics, trade-association material, public procurement notices, and recurring questions from customer conversations.
- Keep the source URL, date checked, and the exact passage or figure that matters.
- Separate direct evidence (a published price or a repeated customer request) from interpretation (why it might be happening).
- Note what is missing: a public price list does not show how many clients buy, and an online review is not a representative survey.
For a first pass, ten carefully chosen sources are usually more useful than a hundred loosely related search results. If you use your own client notes, remove names and identifying details unless your AI setup and purpose are approved for that data.
Use AI as an analyst’s assistant
Give the tool the question and the sources you collected. Ask it to extract claims into a table with columns for source, date, customer segment, evidence, and uncertainty. Then ask for competing explanations and the strongest evidence against its first interpretation. Require citations back to the supplied source, and check each one yourself; a citation-shaped answer is not verification.
AI is useful for grouping repeated phrases, comparing how providers explain similar services, and drafting a short list of follow-up questions. It is much less reliable at estimating market size, forecasting revenue from sparse evidence, or distinguishing an unusual example from a real trend. Keep those judgments with you.
Turn findings into a low-risk test
Finish with one practical experiment, not a grand strategy. You could rewrite one service page for a clearly defined audience, ask five existing customers what nearly stopped them from buying, or test a small offer with a handful of suitable prospects. Set a time limit and a signal in advance: replies, qualified enquiries, or booked conversations. A few responses do not prove the size of a market, but they can tell you whether the next test is worth doing.
When comparing prices, normalise the scope before drawing conclusions. A CHF hourly rate, a fixed-fee package, and a monthly retainer are not directly comparable if they include different work. Record what is included, who the offer is for, and when you checked it. Use your own costs and capacity to set a price; competitor pages are context, not a pricing formula.
Keep the boundaries clear
Use public material in line with its terms and respect robots, access restrictions, and copyright. Do not feed identifiable client correspondence or confidential business information into a general-purpose AI tool without checking the provider, settings, purpose, and your obligations under Swiss data-protection rules. A private research question rarely needs private customer data.
The useful outcome is not “AI says the market is growing.” It is a short, traceable note: what you observed, what remains uncertain, what decision you made, and what you will test next. That is market research a busy independent professional can actually use.
