Client Profile.
The client is a UK based company empowering organizations to move towards process excellence by automating their legacy processes, increase efficiency, and reduce costs. With over a decade’s experience, state of the art technology, expertise, and supporting services, the company is on an aggressive expansion drive to expand its solution offerings, technical partners, customer base and sectors they operate in.
Business Need.
The client was looking to do away with manual, time-consuming and error-prone accounts payable process for its end clients with a fully automated order-based invoice processing system. The technology-driven end-to-end AP process would lead to timely supplier payments through enhanced visibility of invoices in process.
Challenges.
- Manual verification of multiple data fields such as supplier name, invoice date and number, purchase order number, currency etc., extracted through OCR process from scanned invoice images.
- Ensuring accurate entry of data points which were incorrect/missing in scanned invoice copies.
- High volumes of invoices for approvals and archiving increased chances of errors, stretched the approval process and called for larger workforce.
Solution.
A structured invoice management process was developed right from receipt of invoice and matching with purchase order to approval and final posting; all through a single web-based client interface. Automation and macros were extensively used to accelerate the process and increase accuracy. Purchase order-based invoices were automatically matched to PO data, whereas expense invoices were automatically distributed to designated approver for review and approval.
- Reduced paper invoices and improved accuracy and control
- Dashboards to track where invoices are in process, and where bottlenecks occur
- Reports to analyze deviations on order-based invoices to improve master data quality
- Reports and dashboards highlighting time to payment, time to invoice approval, etc.
Approach.
For ensuring data accuracy at 99.5% and processing all the invoices received in a batch and delivering them back within 24 hours; a robust automated workflow was put in place. HitechDigital’s project team collected detailed requirements from Procurement, Finance and IT departments. Accordingly, algorithms and scripts were written to automate the workflow that fit their specific needs.
Receiving input documents from client
- Client saved Invoices in PDF form on the FTP server, and macros sent email alerts with the count of invoices uploaded.
Classification of invoices
- Scanned invoices were classified primarily according to file format, and then as per the line and header items including vendor name and type, invoice number and date, amount and currency etc.
- To segregate invoices according to vendors scripts were written, while macros ensured 100% accuracy in vendor-wise data segregation.
- Reconciling invoices with purchase orders and sending mismatch alert/notification was managed through bots specifically programmed to execute the task.
- Bots also generated daily report about invoices not in standard format and processed and unprocessed invoices.
Data Entry of extracted information
- Algorithms, to send invoices for OCR data extraction, were written to first check if invoices were valid or invalid according to criteria set by departments. Invoices in invalid formats were not processed.
- Workflow would automatically route extracted data for structured entry in the web-based tool.
- Automated bots backed with human intelligence were leveraged to send processed invoices back to the client, accompanied with a detailed report and an email notification.
Quality check of data entered
- Templatized and trained bots performed routine yet critical data validation and looked for missing values if any; by mimicking the way data professionals interact with applications.
- Discrepancies or errors in the data were highlighted with help of macros.
- Rules/algorithms backed with Machine Learning capabilities not only identified errors; but also captured new types of errors.
Business Impact.
Automated Purchase-to-Pay process reduced cost and increased efficiency.
In-house resources redeployed to focus on value adding activities.
Improved supplier/vendor relationship.
Reports and dashboards to enable informed decision making.
