The verdict in three sentences
Trade finance banks, counterparties, candidates and even NGOs now ask AI assistants like ChatGPT about a Geneva commodity trading house before getting in touch, and the answer relies on what is published and consistent about you. A house whose governance, compliance and office pages are thin or contradictory gets described vaguely, inaccurately or not at all. AI visibility work costs, as a 2026 ballpark, CHF 3,000 to 25,000 for a diagnostic followed by an initial project, and it is measured through a dated record of the answers to the same questions, never through a promise of being recommended.
What do a trader's counterparts ask AI assistants?
A trading house does not sell to the public. People who ask an AI assistant about it want to verify, compare or assess a risk. Here are realistic questions, phrased the way they are asked in Geneva in 2026.
| Who asks | Example question to an AI assistant | What the answer needs to find |
|---|---|---|
| Trade finance banker | “Which Geneva-based grain trading houses publish a sanctions policy?” | Dated, accessible compliance page |
| Industrial counterparty | “Does this Geneva metals trader have offices in Asia?” | Up-to-date list of offices |
| Buyer at a producer | “Which Geneva traders publish a responsible sourcing report?” | Report actually published, date, scope |
| NGO or journalist | “Who runs this trading company and how is it governed?” | Management and board page, group structure |
| Analyst candidate | “Which trading houses in Geneva hire junior energy operators?” | Careers page, programmes, working languages |
| A client's ESG auditor | “Does this trader publish a sustainability report?” | Report as published, no extrapolation |
| Credit insurer | “How long has this group been based in Geneva?” | Factual, dated history |
The common thread: none of these questions is about a product. They are about credibility and compliance, which are exactly the pages many traders keep minimal out of discretion.
Why some houses appear in answers and others do not
AI assistants like ChatGPT build their answers from public sources they treat as reliable and consistent. The discretion typical of Geneva trading has a cost: when a company publishes little, the assistant fills the gap with whatever it finds elsewhere, sometimes an old article, a registry entry or a confusion with a similarly named firm.
The factors that come up most often:
- Consistency: the same legal name, offices and executives on the website, registries, professional profiles and industry association directories.
- Third-party mentions: business press, trading associations, publications by banks or industry bodies that cite the company.
- Clear factual pages: one page per topic (governance, sanctions and KYC, responsible sourcing, offices), written plainly and dated.
- Structured data: markup for the organisation, its addresses and its executives, which helps engines connect the information.
- Freshness: a 2021 ESG report left alone online suggests the company has published nothing since.
Care is needed with commitments: a page describing a compliance framework must reflect what actually exists, and its wording deserves review by your legal team, all the more since Switzerland applies its own sanctions measures, implemented by SECO. This is not legal advice. On personal data, the revised Federal Act on Data Protection (revFADP) has applied since 1 September 2023 to the site's forms and analytics tools, and the GDPR may also apply for visitors located in the EU.
What the work covers and what it costs on the market
In broad terms, an AI visibility project for a trader covers a review of what AI assistants say today, alignment of public information, rewriting or creating factual pages, technical markup and tracking over time. The market is young: the ranges below are 2026 ballparks drawn from typical quotes by agencies and consultants in Geneva and French-speaking Switzerland, and should be read with caution.
| Service | Broad content | 2026 ballpark (CHF) | Frequency |
|---|---|---|---|
| Initial diagnostic | Record of answers to 20 to 40 questions, gaps and errors | 3,000 to 8,000 | One-off |
| Information alignment | Website, registries, profiles, industry directories | 4,000 to 12,000 | One-off |
| Credibility pages | Governance, compliance, offices, careers | 6,000 to 20,000 | One-off |
| Structured data and technical work | Organisation, address and executive markup | 2,000 to 6,000 | One-off |
| Answer tracking | Monthly or quarterly dated record | 800 to 3,000 per month | Recurring |
| Trade press relations | Mentions in specialist titles | 3,000 to 10,000 per month | Recurring, optional |
At Kolonell, this work is priced on quote after a first conversation: the scope depends on the number of group entities, the languages and the state of existing pages.
How to measure without fooling yourself
Need a professional website?
Kolonell builds websites that attract clients, optimized for the Sénégalese market. Free quote in 2 minutes.
Measurement rests on a dated record: the same list of questions, put to AI assistants like ChatGPT at regular intervals, with the answers archived. You track the share of answers that mention the company, the accuracy of what is said (offices, executives, activities) and, when the tool shows its sources, whether the company's pages appear among them. Answers vary from one day to the next, so the trend over three to six months is what counts, never a single screenshot.
Mini case study
Illustrative example: Rhône Meridian Trading, a fictional 60-person base metals trader in Geneva, finds that an AI assistant credits it with an office closed in 2022 and ignores its sanctions policy. Management orders a diagnostic on 30 questions.
First record: the company appears in 6 answers out of 30, or 20%, and 4 of those 6 answers contain an error. The first-phase budget, as a market assumption, includes a CHF 5,000 diagnostic, CHF 6,000 of information alignment and three rewritten pages at CHF 9,000, so 5,000 + 6,000 + 9,000 = CHF 20,000. Quarterly tracking will show whether the share of accurate mentions grows; no result is promised, and the decision to continue will be taken on the six-month record.
FAQ
Does a discreet trading house really need to care about AI visibility?
Yes, because discretion does not stop AI assistants from answering: they simply rely on third-party sources, sometimes outdated ones. A handful of well-maintained factual pages, often 4 to 6, is enough to fix the essentials without revealing anything commercial.
How much does an AI visibility diagnostic cost in Geneva?
As a 2026 ballpark, a serious diagnostic costs CHF 3,000 to 8,000 depending on the number of questions and languages. At Kolonell, it is priced on quote.
Can we be sure ChatGPT will cite us after this work?
No. No serious provider can promise that an AI assistant will recommend a company: the work improves the quality and consistency of available information, then measures the change over 3 to 6 months.
Do we need to publish our volumes or counterparties?
No, nothing requires you to publish commercial data in order to be described accurately. Governance, offices, published policies and dates are enough in most cases, and what becomes public remains a decision for management and its counsel.
Which language should we publish in for Geneva?
English is the language of counterparties and banks, while French remains useful for the Swiss French-language press, local authorities and recruiting. A record in two languages usually costs 30 to 50% more than a single-language record, as a ballpark.
Find out how AI assistants describe your firm. A dated record of what AI assistants like ChatGPT say about your house today, with the gaps to fix first. Request a diagnostic quote: reply within 48 hours.
Mohamed Bah
Fondateur, Kolonell
Passionate about digital and entrepreneurship in Africa, Mohamed has been helping Sénégalese businesses with their digital transformation since 2020. Founder of Kolonell, he believes every SME deserves a professional and accessible online présence.
