As of 13 August 2026, AI can draft a purchase order.
This still needs a person who signs their name to it.
Can you do it?
5 minutesto a draft.
30 minutesto something you’d act on.
Cost, all in£0
Skill neededchat-fluent
Who has to check ityou
What the alternative costsProcurement software like Coupa or Jaggr integrates PO generation with supplier management, typically at enterprise cost; a template-based tool or spreadsheet costs nothing.
If this goes wrong: you order the wrong quantity, ship to a stale address, price yourself into a loss, or omit payment terms, and discover the error after the supplier has fulfilled the order.
What to actually do
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.
Use a tool built for this
Second choiceDo it yourself
The distant thirdA chat interface, chat-fluent skill, and roughly 30 minutes until you can act on the result.
How to actually do it
- Gather the information you need: your company details, the supplier's full name and address, product codes, quantities, unit prices, the delivery address if different from your company, and your payment terms.
- Open ChatGPT, Claude or Gemini and paste the prompt above, filling in the bracketed placeholders with your actual details.
- Read the draft and check every number, address and product code against your source documents (price list, supplier record, delivery instructions).
- Verify that the supplier address is current by checking your latest supplier record or recent invoice; AI may use stale information.
- Add or amend any special clauses your company requires (quality standards, inspection rights, returns policy) and paste the revised text back into the chat to regenerate if needed.
- Copy the final version into your company's PO system or email template, assign your PO number if the AI did not, and send to the supplier through your normal process.
- Keep a copy with your purchase record and match it against the supplier's acknowledgment and later invoice.
Prompt
Draft a UK purchase order using the details below. Include our purchase order number [YOUR PO NUMBER], order date [DATE], and our company details [YOUR COMPANY NAME, ADDRESS]. Supplier: [SUPPLIER NAME AND ADDRESS]. Items: [LIST EACH ITEM WITH PRODUCT CODE, DESCRIPTION, QUANTITY AND UNIT PRICE]. Delivery address: [DELIVERY ADDRESS]. Payment terms: [E.G. NET 30]. Include standard clauses for goods receipt, invoice matching and our right to cancel if delivery is late. Use formal tone and format it as a single-page document ready to send. Do not invent product codes, prices or supplier details.
Open it prefilled in ChatGPT or Claude, or copy it into Gemini, which takes no prefill link.
What it gets wrong
- Cannot verify that supplier addresses, bank details or product codes are current without checking them independently against your own records.
- Invents plausible-sounding product codes or descriptions if you omit them, so every code and specification must come from you or your price list.
- Cannot know your company's legal standard terms, exclusions or liability caps; you must add these yourself or your default terms are absent.
- Does not flag when a unit price is unusually high or low, so you must sanity-check the commercial terms before committing.
- The order is your obligation once sent; the model's liability is zero.
Even on a YES, the friction has a name: stakes of error, verification cost and legal accountability.
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.
| Axis | Score (0–2) |
|---|---|
| Output | 2 |
| Inputs | 2 |
| Verification | 1 |
| Liability | 1 |
| Effort delta | 2 |
| Total | 8 / 10 |
The methodology and its thresholds are published in full.
FAQ
- Can AI draft a purchase order I can send straight away?
- Almost. It produces a structurally correct document with all the standard fields, but you must verify every number, address and product code against your own records first. The failure is costly: a wrong quantity or stale supplier address means you discover the error after the goods are ordered.
- Will AI invent product codes or prices?
- Yes, if you do not supply them. Always provide real product codes, descriptions and unit prices from your own systems; do not let the model fill gaps. If you give it incomplete information, it will plausibly invent the rest.
- What happens if the purchase order goes wrong?
- You are liable. Your signature and company name are on it, so if the supplier has fulfilled an order with a wrong address, quantity or price, that is your obligation to sort. The model has no liability; you do.
- Is there software that does this better than chatbots?
- Yes, if you have a supplier database. Procurement platforms integrate PO generation with your supplier records so codes and addresses are pulled from a live source, not re-entered each time. For occasional POs without that infrastructure, a chatbot and a template are enough.
Nearby answers
Assessed by claude-haiku-4-5 (claude-haiku-4-5-20251001) 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.