Reading emails for Cashflow automation

Email is a well-established and widely adopted tool for information exchange, having been in continuous use since the early days of the Internet. Its ubiquitous capabilities also make it a straightforward tool for automating the constant flow of cashflow information customers produce.

A typical email-based automation workflow starts when the customer provides a dedicated address for remittance advice. Suppliers must send all payment documents to this email to ensure proper processing.

Thanks to the flexibility of the email format, the remittance information can be provided in many ways. The customer can embed it in the mail body or send it as PDF, XLSX/X, CSV, or image attachments, for example.

This flexibility, while beneficial, also introduces significant challenges for finance staff. Emails can arrive in virtually any format—plain text, HTML, with or without attachments—requiring robust and adaptable parsing mechanisms for automation.

To address this, Cashbook engineers have implemented a detailed parsing process that examines each incoming email to identify and extract the relevant remittance information provided by the customer’s supplier. This part is the most delicate part of the process. We thoroughly analyze the email, searching for table information in the body or text. If we find nothing there, we then process each attachment.

There is a custom processor for each one of the most common formats the attachments can present. That includes having a format processor for spreadsheets, another one for CSV files, same for plain TXT files, for images like TIFF/ TIF/ GIF/ JPG, MS Word documents, etc. Previously we required templates by customer but through years of experience & machine learning we have eliminated this requirement. 

Sometimes the customer presents a very specific scenario, and they might request a custom format for its specific way of supporting the information provided.  This is very common in businesses with deduction claims.  We call these Fixed Formats.

We identify the proper format processor, including any custom ones. Then, we extract information using its specific algorithms. Generally, this means reading the attachment or mail body for structured data. Next, we build a data table and convert it into an intermediate format.

Finally, Cashbook will read and interpret that intermediate format, no matter how it was generated from, and it will carry on with the automated operations like data storage, data verification (when needed), allocation algorithm running, and eventually linking the remittances received to the bank files imported. This closes the Cashbook Email Cashflow Automation process.

One of the additional benefits of this type of automation above and beyond time savings is  the use of these remittances for Credit & Collections staff to review payment history.