> ## Documentation Index
> Fetch the complete documentation index at: https://docs.upflow.io/llms.txt
> Use this file to discover all available pages before exploring further.

# AI-suggested promise to pay

> Let Upflow AI scan customer replies and invoice context to suggest a realistic promise-to-pay date your AR team can confirm in a single click.

Upflow can now automatically suggest a promise to pay date when a customer indicates they'll pay later. This feature analyzes customer communications and invoice context to recommend the most suitable payment timeline.

## When is this useful?

This feature is helpful when:

* You're managing multiple payment conversations and need quick, data-driven recommendations
* You want to maintain consistent follow-up practices across your team

## Prerequisites

To use AI-suggested promise to pay:

* **Email integration**: You must have [automatic reply imports](/en-us/collection-and-collaboration/emails/import-your-customers-replies-automatically) configured so Upflow can analyze customer communications

## How it works

Upflow analyzes customer communications (e.g., email replies imported into Upflow) to detect when a customer is **committing to pay later**. When a promise is detected, Upflow:

1. **Identifies the invoice(s) concerned** based on the message context and your customer/invoice data.
2. **Suggests a Promise to Pay date** based on the conversation
3. **Surfaces the suggestion in-context**, along with the supporting excerpt from the customer message when possible.
4. **Keeps you in control**: you can **accept** the suggested date, **edit** it, or **reject** it and set a different date.
5. **Logs the outcome** once confirmed so your team can follow up consistently and track promises over time.

<img src="https://mintcdn.com/upflow-04dc27cd/qndk5R7KgkVa_zPO/images/44395784011153.png?fit=max&auto=format&n=qndk5R7KgkVa_zPO&q=85&s=88e07f96c63a8ba1f102a0e8589a33a0" alt="" width="640" data-path="images/44395784011153.png" />

The promise to pay that is suggested will include a quote from the communication that explains the rationale behind the suggestion. You can always edit it.

<img src="https://mintcdn.com/upflow-04dc27cd/qndk5R7KgkVa_zPO/images/44395804402321.png?fit=max&auto=format&n=qndk5R7KgkVa_zPO&q=85&s=0e9532520cf5c131305269d11dd952fb" alt="" width="472" data-path="images/44395804402321.png" />

## Limitations

* AI suggestions are based on imported email content only - they cannot analyze calls, letters, or conversations outside Upflow
* The quality of suggestions depends on clear communication in customer emails
* Manual review is always recommended to ensure the suggested date aligns with any specific agreements or circumstances

## Frequently asked questions

**Can I override the suggested date?**

Yes. The AI provides a recommendation, but you have full control to accept, modify, or ignore the suggestion and enter your own date.

**What happens if I don't agree with the suggestion?**

Simply enter the date you prefer. The AI learns from actual outcomes over time, so your corrections help improve future suggestions.

**Does this work for all invoices?**

The feature works best when there's sufficient customer history and clear communication about payment intent. For brand new customers or ambiguous replies, suggestions may be less precise.

**Is this available for all integrations?**

Yes, as long as your accounting system is connected to Upflow and automatic reply imports are configured, the feature works regardless of your integration type.