$ 2.4B leading healthcare network and clinical care provider in the United States operating 7 hospitals and 50+ outpatient/urgent care centers. The client has more than 3,000 affiliated physicians in all specialties including many with a national and international reputation.
The Accounts Payable department had more than 25 employees dedicated to invoice processing, with a monthly invoice volume of approximately 40,000. The current process was very manual, labor-intensive and prone to errors, facing the typical challenges associated with any AP processing organization with lack of automation:
- Lost or missing invoices
- Slow invoice approval times
- Discrepancies and exceptions
- High turnover, impacting productivity and quality
- Lack of visibility into outstanding liabilities and overall performance metrics
- No time available for pro-active vendor management
Through a recent acquisition, the organization almost doubled in size and the processing volumes had rapidly increased to 70,000 invoices per month. As part of the growth, the Finance organization had the mandate to absorb the additional AP transactional volume with a minimal increase in staffing levels and operating costs.
In order to achieve this challenging expectation, Auxis’ Intelligent Automation experts were brought in to analyze the current AP operation, find opportunities for automation, and propose a more scalable model for the organization going forward.
Auxis started the engagement by setting realistic expectations and explaining to the client that they should not anticipate achieving 100% “touchless” automation. Instead, Auxis helped the organization establish a more realistic target of the percentage of invoices that could become “touchless”, which is typically between 60% to 80%.
“Auxis designed and implemented a new, optimized AP process that incorporated Robotic Process Automation (RPA) with UiPath and Intelligent Data Capture”
As part of this analysis, Auxis performed a detailed breakdown of invoices by supplier, identifying the top 120 vendors that represented 70% of the total invoice volume. Auxis also identified that 90% of the invoices had a related PO, making the current invoice processing step even more time-consuming. This process required a 3-way match which had to be manually done line by line. In order to automate the processing of these PO invoices, the client needed a solution that could turn unstructured data from vendor PDF invoices into structured data at the line item level and not just at the header level.
With all of these insights available, Auxis designed and implemented a new, optimized AP process that included Robotic Process Automation (RPA) and Intelligent Data Capture technology.
The RPA software implemented was UiPath, and the IA Data Capture software was Ephesoft. These technologies were integrated into two main processes:
- Invoice Indexing: Utilizing the AI Data Capture platform “Ephesoft Transact”, Auxis automated the extraction of the header level data for all vendors, and the line by line data for the top vendors previously identified. The new automated step-by-step process included:
- Receiving the AP documents through email and scanner
- Classifying the received documents by type:
- Vendor Statements
- Invoice – Header Extraction
- Invoice – Line by Line Extraction
- Credit Memos
- Identifying the associated vendor by searching on the vendor master data matching fields such as phone number, remit to address, and incoming email domain. For unidentified vendors, Ephesoft places the document in a queue for staff review to either update the vendor master or add a new vendor.
- Extracting the required header information from the invoice for each document type
- Extracting the line by line data for the high-volume vendors
- Saving the document to the Document Management system and indexing with the header information for all vendors
- Sending information to the next process (voucher creation) for invoices that require manual creation of a voucher
- Invoice Processing: Leveraging the RPA UiPath platform, Auxis also automated the voucher creation in the ERP system (PeopleSoft) as well as the three-way match process. In this case, the new step-by-step process performed by the robot included:
- Receiving the invoice detail from the Ephesoft platform, including header information and detailed table extraction (line by line)
- Searching the associated PO in the system and matching with the right vendor
- Matching the invoice lines to the PO lines including item number, quantity, price, etc.
- Validating and identifying discrepancies
- Taking action to adjust line item level discrepancies. For example, for line items in the invoice that are not in the PO, the robot adds those new lines and routes for required approval or denial.
- Once the robot has completed processing each batch, it creates a “processing log” that contains each item processed by the robot, including the outcome for each invoice (successful or unsuccessful).
- Lastly, the robot saves a copy of the invoice and the processing log in a shared folder, and send it via email to the AP staff so they can manually adjust and process the unsuccessful invoices.
With Auxis solution in place, the client was able to achieve 66% touchless invoice processing, equivalent to freeing up over 2,300 human hours per month.
In addition to freeing up capacity and allowing employees to focus on more rewarding, higher-value activities, the new automated process also brought the following benefits:
- Invoice backlog reduction from four days to less than 4 hours. In essence, there are no pending invoices to be processed, just the ones that are received and entered on the same day.
- The backlog reduction translated into increased visibility to the total outstanding liabilities.
- Improved data accuracy and controls with the elimination of manual data entry.
- Less dependency on the staff in order to get the job done. The turnover became a less important challenge and is expected to go down as now employees are not overwhelmed with processing transactions and can focus their time on more interesting activities.