The verdict in three sentences
A commodity trading house in London sells little through its website, but it is judged on it: trade finance banks, counterparties, NGOs, regulators and recruits use it to check governance and compliance. These audiences now put their due diligence questions to AI assistants like ChatGPT, which answer from whatever consistent information they find on the web, or from nothing. The main stake is accuracy: what the assistant says about your offices, your sanctions policy and your responsible sourcing must match what you actually publish.
Who asks AI assistants about a trader, and what do they ask?
The questions below are realistic phrasings, typical of the different audiences of a trading house. None of them is about price or product.
| Audience | Question put to an assistant | What the answer needs to find |
|---|---|---|
| Trade finance bank | "Which mid-sized metals traders are headquartered in London and publish a sanctions policy?" | Dated compliance page, legal entity, head office address |
| Mining supplier | "Does this London trader have a responsible sourcing policy for cobalt?" | Published policy, standards followed, last update date |
| Energy counterparty | "Who runs the oil products desk at [company] and where are its offices?" | Leadership page, list of offices |
| NGO or journalist | "Does [company] publish an ESG report, and with what scope?" | Report actually published, exact scope, year |
| Graduate trader | "Which trading houses in London hire junior analysts in agriculture?" | Careers page, programmes, locations |
| Partner's KYC reviewer | "Is [company] a subsidiary of a group and which countries does it operate in?" | Group structure, countries of operation |
| Debt investor | "Which London trading houses publish detailed annual accounts?" | Link to public filings, governance |
| Experienced hire | "How does [company] compare with the large traders on culture and pay?" | Values, career structure, no invented figures |
The common finding: on half of these questions, a mid-sized trader does not appear, or appears with outdated information (old office, departed executive, previous year's report).
Why some houses appear and others do not
Assistants do not "pick" a trader by size. They reuse information that is public, consistent and easy to attribute. The factors that weigh most, in general terms:
- clear factual pages on the entity, offices, leadership and business lines, without marketing jargon;
- compliance policies (sanctions, KYC, responsible sourcing, ESG reporting as actually published) that are accessible and dated;
- third-party mentions: public registers, specialist press, industry associations, NGO reports;
- structured data describing the organisation and its locations;
- freshness: a 2023 report still presented as the latest erodes the confidence of the assistant and of the reader alike.
One constraint specific to the sector: a trading house should never publish volumes or counterparties it does not already disclose officially. The work makes what exists legible, without adding numbers.
On the UK framework, the website itself falls under UK GDPR and the Data Protection Act 2018, overseen by the ICO, and under PECR for cookies and marketing email. If a group entity is FCA-authorised, any financial promotion follows FCA rules. The Equality Act 2010 makes WCAG 2.2 AA the sensible target. This is not legal advice.
What the work covers, and what it costs on the market
| Component | Scope, in broad terms | 2026 ballpark (London) |
|---|---|---|
| Initial diagnostic | Record of current answers to 25 to 40 questions, gaps against published facts | 3,000 to 8,000 GBP fixed fee |
| Upgrade of factual pages | Entity, offices, leadership, compliance, careers | 5,000 to 18,000 GBP depending on volume and legal review |
| Structured data and external consistency | Markup, alignment of profiles and public registers | 2,000 to 6,000 GBP |
| Monthly monitoring | Dated record, corrections, watch on third-party sources | 2,000 to 6,500 GBP per month |
| Internal compliance review | Time from your legal and compliance teams | Internal cost, often underestimated |
| Translations (if offices in Geneva, Singapore, Dubai) | Third-party provider | Depends on languages and volume |
These ranges reflect typical London agency quotes and public agency rate pages; the market is young and gaps between quotes remain wide.
Measurement rests on a dated record: the same questions asked every month, answers archived, with three indicators: the share of answers citing the house, the accuracy of what is said, and the sources drawn on. No serious provider promises that an assistant will cite or recommend your company; it commits to verifiable work and honest measurement.
Need a professional website?
Kolonell builds websites that attract clients, optimized for the Sénégalese market. Free quote in 2 minutes.
Mini case study
*Illustrative example.* Thameside Commodities, a base metals and concentrates trader with 140 staff and offices in London and Geneva. Out of 32 questions recorded in February, the company is cited in 6 answers (around 19%), 4 of which still mention a closed office and a 2023 ESG report.
The head of communications prices the first year: diagnostic 5,000 GBP, page upgrade 12,000 GBP, monitoring 3,000 GBP per month for 10 months, so 5,000 + 12,000 + 30,000 = 47,000 GBP. He weighs it against the cost of a single KYC query delayed at a bank: several days of compliance team time. The goal set: by the December record, answers citing the company describe the right offices and the right report.
FAQ
Does a trading house really need visibility in AI assistants?
Yes, because its audiences (banks, counterparties, recruits) increasingly check through an assistant before a first conversation. It is mostly about accuracy: out of 30 questions, a single wrong answer on sanctions can cost more than a year of monitoring.
How much does this work cost in London?
As a 2026 ballpark, a diagnostic of 3,000 to 8,000 GBP, then monitoring of 2,000 to 6,500 GBP per month. The internal cost of legal review comes on top and should be planned.
Should we publish more trading figures?
No: nothing should go beyond what the company already discloses officially. The work is about clear, dated pages consistent with the 5 or 6 third-party sources that matter.
How is progress measured?
Through a monthly dated record of answers to the same list of 25 to 40 questions. Track the share of answers citing the house and the accuracy of the facts they repeat.
Does this replace press relations?
No, the two work together: mentions in the specialist press are among the sources assistants draw on. A good diagnostic shows which ones are missing.
Find out how AI assistants describe your firm. A diagnostic records what assistants say about your trading house today and the gap with your official disclosures. 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.
