Need a hand?

Just pop your question below to get an answer.

Approval recommendations

What are approval recommendations?

If you're on a Grow, Scale, or Enterprise plan, you can use the AI approval recommendations to bring your company’s spend rules directly into your Expense approval process.

Instead of manually checking every receipt against the policy document, approval recommendations will provide instant suggestions for every expense approval request.

Approval recommendations currently supports approvals for 'Card expenses' and 'Reimbursements'.

Set up approval recommendations

You first need to upload and publish your company's expense policy document. You'll need the 'Manage Business' permission to continue (located under 'Platform' and 'Billing').

  • Go to 'Security & controls' and then to the 'Policy' section
  • Click 'Upload Policy' and select your company’s spend policy document (PDF or Word). Alternatively, you can click 'Add manually' to copy paste, or type your expense rules
  • Click 'Publish' when you're ready. Our AI will begin processing the document. This might take a few minutes. Once finished, all new incoming expense requests will automatically be assessed by approval Recommendations

Understanding approval recommendations

Once your policy is active, approvers will start seeing intelligent recommendations within 'Requests' for card expenses and reimbursements.

  • Approval recommended: the AI has determined the expense complies with your uploaded policy
  • Reject recommended: the AI has identified a specific rule violation or missing information (for example, an exceeded spend limit, missing receipt, or amount mismatch)
  • Manual review recommended: the AI has found ambiguity in the policy or receipt and suggests a human review

Important notes

  • Approval recommendations are provided to help you decide. They don't automatically approve or reject requests. A human approver always makes the final decision
  • Approval recommendations use your expense policy rules as the main source of truth. The more detailed and clear your policy is, the more accurate the recommendations will be
  • Data privacy: your policy data is used exclusively for your account’s recommendations. It isn't used to train models for other customers