Home · Business · Finance & Accounting · Invoicing & credit control

PARTLY

As of 13 August 2026, AI can only partly send your invoices automatically.

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

Can you do it?

30 minutesto a draft.

2 hoursto something you’d act on.

Cost, all in£0/month

Skill neededpower-user

Who has to check ityou

What the alternative costsA purpose-built bookkeeping automation tool is the alternative; no price is provided here.

If this goes wrong, an incorrect or duplicate invoice reaches a customer and creates collection, cash-flow or accounting work for your business.

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, power-user skill, and roughly 2 hours until you can act on the result.

    How to actually do it

    1. Open your current invoicing or accounting system and export the customer list, product or service price list, VAT settings, payment terms and billing schedule.
    2. Remove unnecessary personal data, then gather the exact fields the system uses for customer email addresses, invoice numbers, dates, descriptions, prices, VAT and due dates.
    3. Paste those fields and the system name into the prompt, and ask the chatbot to produce the mapping, stop conditions, templates and test plan.
    4. Create the workflow in your invoicing or automation system using a recurring billing trigger, and add a manual approval step for new customers, changed prices, unusual VAT treatment, missing data and duplicates.
    5. Run the workflow with test customers and test email addresses, then compare each generated invoice with the current price list, customer record, payment terms and accounting entry before enabling sending.
    6. Enable automatic sending only for the approved routine cases, keep the sending log, and check the first live invoices against the accounting records and customer replies.
    7. Ask an accountant to check the setup if the invoices involve complex VAT, foreign customers, credit notes, construction industry rules or disputed amounts; this is not professional advice.

    Prompt

    Design a safe setup for sending my routine UK business invoices automatically. Use only the information I provide and invent nothing. Business details: [business name, address, company or sole-trader status, VAT status]. Customers and billing rules: [customer names or IDs, email addresses, products or services, prices, VAT treatment, billing dates, payment terms, currencies]. Accounting or invoicing system: [system name and available integrations]. Sending method: [email account or automation platform].
    
    Produce:
    1. A field-by-field mapping from my source data to the invoice fields.
    2. The recurring trigger and the exact conditions that must be met before an invoice is created or sent.
    3. A list of cases that must stop and require manual approval, including missing data, changed prices, unusual VAT treatment, duplicate invoices, failed payments and new customers.
    4. A test plan using test customers and test email addresses before any real invoice is sent.
    5. A plain email template for the invoice and a separate payment reminder.
    6. A checklist for comparing the generated invoice with my current price list, customer record, payment terms and accounting records.
    7. The information or access still needed from me.
    
    Do not send anything, make tax judgements, guess VAT treatment, or claim that an integration exists unless I have confirmed it. State clearly where an accountant should check the setup. This is not professional advice.

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

What it gets wrong

  • AI cannot access your live customer, accounting and payment records unless you connect systems and grant the required permissions.
  • AI cannot decide whether an unusual transaction has the correct VAT treatment or whether a credit note is appropriate.
  • AI cannot recognise every commercial exception, such as a changed contract, disputed delivery or customer who should not yet be billed.
  • AI cannot carry your business's liability when an invoice is wrong, late, duplicated or sent to the wrong recipient.
  • AI cannot replace an accountant's review of a serious or complex invoicing and VAT setup.

What caps this at PARTLY: legal accountability, verification cost and private data access.

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
Inputs1
Verification1
Liability1
Effort delta2
Total7 / 10

FAQ

Can ChatGPT send invoices automatically?
Not by itself. A chatbot can design the workflow, draft invoice data and produce setup instructions, but an invoicing or automation system with access to your records must create and send the invoices.
Is it safe to automate sending invoices?
It is suitable for routine, stable invoices after testing and approval rules are in place. Keep manual approval for changed prices, new customers, unusual VAT treatment, duplicates and disputed work because your business remains responsible for errors.
Can AI send VAT invoices for my business?
AI can help populate and send VAT invoices through a connected invoicing system, but it cannot reliably decide the correct VAT treatment for every transaction. This is not professional advice, and a serious or complex VAT case needs an accountant.
What do I need to automate invoice emails?
You need accurate customer email addresses, products or services, prices, VAT settings, billing dates, payment terms and an invoicing or accounting system that supports automation. You also need a test process, stop conditions and a way to check sending logs against your accounting records.

Nearby answers

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.