PARTLY

As of 13 August 2026, AI can only partly check whether a supplier is genuine.

This still needs a person who signs their name to it.

Can you do it?

5 minutesto a draft.

1 hourto something you’d act on.

Cost, all in£0

Skill neededchat-fluent

Who has to check ita colleague

What the alternative costsThe supplied tool data gives no price for a human supplier-verification alternative.

If this goes wrong, you may pay a fraudulent supplier, disclose business information or stop an order after relying on a convincing but false report.

What to actually do

  1. Hand it to a person

    The route this page recommends

    A person who owns the outcome does this end to end, worth it when the failure is dear.

  2. Use a tool built for this

    Second choice

  3. Do it yourself

    The distant third

    A chat interface, chat-fluent skill, and roughly 1 hour until you can act on the result.

    How to actually do it

    1. Open the supplier's quotation, website and correspondence, then copy the legal name, company number, registered address, telephone number, email domain, website URL, payment instructions and proposed order into a working document.
    2. Redact passwords, full bank details, identity documents and unnecessary personal data, then gather the quotation, terms, certificates, references and any previous correspondence that you are allowed to share.
    3. Check the supplier's legal details on the official Companies House register and record whether the name, company number, registered office, status and filing history match the documents you received.
    4. Open the website using the typed address rather than an email link, inspect its contact and terms pages, and independently find a telephone number or email route instead of relying only on the message that introduced the supplier.
    5. Paste the redacted evidence and the prompt into a chatbot, asking it to produce separate sections for verified facts, gaps, inconsistencies, red flags and next checks.
    6. Compare every factual statement in the report with the source documents and official records, then ask a colleague to repeat the key contact and payment checks through independently found details.
    7. Pause payment and ordering while material inconsistencies remain, and send the supplier specific questions about the legal entity, payment account, delivery terms and references before making a decision.

    Prompt

    Assess whether this supplier appears genuine for a UK business. Do not claim that the supplier is genuine or fraudulent. Separate verified facts, unverified claims, inconsistencies and sensible next checks. Compare the supplier's legal name, company number, registered address, website domain, email domain, telephone number, quotation, payment details and trading history where information is provided. Tell me which checks I can make through official UK sources, such as Companies House, and which facts require contacting the supplier or an independent contact route. Do not treat a website, company registration, positive review or supplied document as proof on its own. Do not infer that a bank account belongs to the supplier from its sort code or account name. Flag requests for payment to a different legal entity, urgency, unusual payment terms, copied wording, mismatched addresses, newly changed bank details and refusal to provide reasonable evidence. Do not ask me to paste passwords, full bank details, identity documents or other unnecessary personal data. End with a cautious recommendation: proceed, pause for checks, or reject the supplier, explaining what evidence is still missing. Supplier information: [paste redacted details, documents and URLs here]. My purchasing context and proposed order: [describe here].

    Open it prefilled in ChatGPT or Claude, or copy it into Gemini, which takes no prefill link.

What it gets wrong

What caps this at PARTLY: stakes of error, judgement under ambiguity and verification cost.

How we scored this

Five axes, each scored nought to two by hand: ten means AI carries the task cleanly, and the thresholds that turn a total into YES, PARTLY or NO are published in the methodology. Each axis name links to its definition.

AxisScore (0–2)
Output2
Inputs2
Verification1
Liability1
Effort delta1
Total7 / 10

The methodology and its thresholds are published in full.

FAQ

Can ChatGPT check if a supplier is legitimate?
It can organise evidence, compare supplied details and identify checks or inconsistencies, but it cannot prove that a supplier is legitimate. Check the underlying facts through official sources and independent contact routes before ordering or paying.
How do I check if a UK supplier is genuine?
Compare the quotation and correspondence with the supplier's Companies House record, website and independently found contact details. Confirm the legal entity, delivery terms and payment instructions directly, and pause if the details do not match.
Can AI check a supplier's bank details?
AI can flag differences between payment details and other supplier information, but it cannot establish who controls a bank account. Confirm new or changed payment instructions through a trusted contact route and do not rely on an AI assessment alone.
Is it safe to use AI for supplier due diligence?
It is suitable for organising redacted information and preparing a list of checks, not for making the final decision by itself. This is not professional advice; a serious suspected fraud or high-value transaction needs a solicitor or fraud specialist.

Nearby answers

All Suppliers & ordering rulings

Assessed by gpt-5.6-luna (gpt-5.6-luna) on 2026-08-13, second-checked by an independent model. Wrong somewhere? Email [email protected] and it gets re-checked.

The newsletter

AI news, new answers and product picks, straight to your inbox.