Home · Business · Finance & Accounting · Bookkeeping & reconciliation

YES

As of 13 August 2026, AI can code your expenses to the right nominal accounts.

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 ita professional

What the alternative costsBooke AI is a purpose-built bookkeeping automation tool for categorisation and reconciliation; no price is stated here.

If this goes wrong: expenses are posted consistently to the wrong accounts or VAT treatment, and correcting the records may affect management reports, VAT returns or accounts preparation.

What to actually do

  1. Hand it to a person

    The route this page recommends

    Someone with a licence or accountable authority has to sign this before it counts.

  2. Use a tool built for this

    Second choice
  3. Do it yourself

    The distant third

    A chat interface, chat-fluent skill, and roughly 30 minutes until you can act on the result.

    How to actually do it

    1. Open your bookkeeping system and export the unposted expense transactions with their references, dates, suppliers, descriptions, amounts and receipt links.
    2. Open the current chart of nominal accounts and copy the account codes and names, including any notes about restricted or commonly used accounts.
    3. Gather the business context, accounting policy and examples of previously approved postings, especially for equipment, travel, entertaining, subscriptions, mileage, home working and director or employee expenses.
    4. Remove unnecessary personal data from the export, then paste the chart, context, policy and transactions into the prompt in the labelled sections.
    5. Ask the model to return the proposed account, confidence, reason and REVIEW status for every transaction, then paste its table into a working spreadsheet beside the original transaction data.
    6. Compare every proposed account with its receipt, business purpose and previous approved treatment, and send all REVIEW items and low-confidence items to your accountant or bookkeeper before posting.
    7. Post only the approved account codes in your bookkeeping system and retain the prompt output with the transaction records as a draft review trail.

    Prompt

    You are helping with UK bookkeeping, not replacing an accountant. Code each expense below to the most appropriate nominal account from the supplied chart of accounts. Use the business context and accounting policy supplied. Do not invent accounts, VAT rates, suppliers, dates or amounts. If an item is ambiguous, unusual, personal, capital, mixed-use, likely to involve a director or employee, or could have a different VAT treatment, mark it REVIEW rather than guessing. For each item, return: transaction reference, suggested nominal account code, account name, VAT treatment only if it is explicitly supported by the supplied information, confidence as high, medium or low, and a short reason. Then list all REVIEW items with the exact missing information needed to decide them. Treat the suggestions as draft bookkeeping only and state where an accountant or bookkeeper should confirm the treatment.
    
    Business context:
    [Describe the business and what it sells]
    
    Accounting policy or previous treatment:
    [Paste any relevant policy or examples of approved postings]
    
    Chart of accounts:
    [Paste account codes and names]
    
    Expense transactions:
    [Paste a table with transaction reference, date, supplier, description, amount, currency, receipt details and any VAT information]
    
    Return the results as a table that can be pasted into a spreadsheet.

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

What it gets wrong

  • AI cannot infer the correct treatment when the receipt does not state the business purpose or when an expense is partly personal.
  • AI cannot reliably settle unusual cases involving capital expenditure, directors, employees, entertaining or mixed VAT treatment without the relevant policy and facts.
  • AI cannot take responsibility for errors in your bookkeeping, VAT return or accounts.
  • AI cannot replace a consistent chart of accounts and an accountant's judgement on transactions that do not fit existing categories.

Even on a YES, the friction has a name: judgement under ambiguity, 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.

AxisScore (0–2)
Output2
Inputs2
Verification1
Liability1
Effort delta2
Total8 / 10

FAQ

Can ChatGPT code my business expenses?
Yes, it can suggest nominal accounts for routine expenses if you provide your UK chart of accounts, transaction details and accounting policy. Treat the result as draft bookkeeping and have ambiguous, VAT-sensitive or unusual items checked before posting.
Can AI choose the right nominal account for VAT?
AI can flag a likely treatment when the receipt and your records contain the necessary facts, but it should not guess a VAT treatment. You or your accountant must confirm the evidence, VAT status and treatment before the transaction is used in a VAT return.
Do I need to give AI my chart of accounts?
Yes. Without your current chart of accounts, the model may suggest account names that do not match your bookkeeping system or your reporting structure. Supply the codes and names, together with examples of how your business has treated similar expenses.
Is AI bookkeeping safe for my business?
It is useful for routine categorisation, but it is not professional advice and it does not carry the liability for your records. An accountant or qualified bookkeeper should check unusual or material items and any treatment that affects a VAT return or statutory accounts.

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.