Automation Potential in: Invoice Processing
low Potential to Automate:
Turn handwritten documents and PDFs into digital data, Chase invoice discrepancies
Medium Potential to automate:
High potential to automate
Turn scanned documents & PDFs into digital data, Collating invoice data from files, Submitting transactions for processing
today's Challenges
According to the Billentis Report, 55 billion invoices are already being sent out worldwide in paperless form…but that’s only 10% of the total. There are significant benefits to the digitisation and automation of invoice processing, which can be carried out faster and without human error.
For almost all companies, invoice processing is an area of significant effort which could be reduced. Accounts Payable spend a large amount of time copying invoice data from a wide range of unstructured documents into an ERP system. This process is prone to human error, due to the high volume of invoices. This isn’t the only challenge:
Manual Entry
Paper invoices either need to be entered manually or scanned in. This incurs considerable additional expense for the recipient and results in errors. Scanning solutions can be error-prone, too. The invoice may be difficult to read, information may be incorrectly attributed etc. Paper invoices bring with them management and space issues of course, but digital invoices are not the sole solution.
Attributing Invoices
Invoices often cannot be clearly attributed to the corresponding orders. The reasons for this can be many and varied: suppliers and purchasing companies may have an invoice ecosystem of their own, using different descriptions and special invoice or article numbers. Sometimes the corresponding information is missing altogether. The supplier might combine several orders in a single invoice, or only issue a partial invoice for a larger order. All of these issues mean that there is a large amount of time spent by AP staff going over documents and matching Purchase Orders and Purchase Request to spot the difference.
Invoice Status
There is often a lack of transparency as to the status of an invoice for both suppliers and customers alike. Has it been received, has it been checked, when will it be paid? The lack of clarity can mean unnecessary time spent by both parties chasing each other.
Slow invoice mailing
Even in this digital age, there are still a lot of cases in which paper invoices are required or preferred. Sometimes days can pass between the time the invoice is created and when it finally gets mailed out to the customer.
There are strong use cases for the automation of the entire invoicing process.
Automation Impact Evaluation
Digitising documents
OCR (Optical Character Recognition) enables organisations to take scanned documents and PDFs and digitise that data. There have been some real advancements in this area in recent years and with a higher degree of accuracy now possible. OCR has yet to fully crack handwritten documents so although some form of digitisation is possible this is not yet at a level where it can handle these types of documents consistently.
Collating invoice data
Accounts Payable spend a lot of time collating invoice data from a range of sources into a central system. Regardless of how the file format (word, excel online) or how that particular file is set out, automation can process it. Robotic Process Automation interacts with files in the same way as people do. It can go into emails and extract documents and will be trained on which data to extract. One of the stages in this process is the setting up of aliases. Where a piece of information may be called different things (for example: Invoice number and Invoice no.) the bot will be trained to find all variations of this information. Once a bot has found the relevant information, it then fills in an ERP and validates the data nearly instantly.
Submit transactions/updates
With the information extracted and validated the bot can then submit transactions to take place, it can also email all the relevant people with updates as to the status of the invoices which should hopefully stop vendors chasing. Where information is missing or there are discrepancies it will or will flag this up for a team member to chase.
case Study
Construction Company
Challenge Faced: This large construction company needed to process over 350,000 invoices a year. Although it was currently using OCR technology Readsoft, it was looking for further ways to increase accuracy, reduce human errors and improve overall time spend processing invoices.
Solutions implemented: Proservartner recommended that the company replace it’s Readsoft OCR which typically had an accuracy rate of around 70% with a combination of Automation Anywhere RPA and IQBOT. With this combination not only were the results of the extracted data more accurate but could be automatically used/matched by the bot. One of the first processes we automated was the search for and elimination of duplicate records which had a significant cost-saving associated with it.
The results: We were able to automate around 80% of the workflow and the company saw a return on investment within 12 months. There was a significant improvement in supplier order being delivered due to faster invoice payment.