One of the biggest obstacles for any organization is consolidating and processing the receipts and invoices for various vendors and suppliers. Organizations that deal with different suppliers and vendors need to also ensure that their payments are processed on time with accuracy. As large amount of paperwork has to be reviewed, data entry carried out, there can be multiple inefficiencies while processing the payment. Late payment can also become a major obstacle. With Straight Through Processing (STP), all such challenges can be addressed.
STP enables direct processing of documents of the supplier to the accounts payable team without any human intervention. This is achieved through our AI-driven solutions that has the ability to identify and extract data intelligently. STP is easy to integrate with your existing solutions can automate your entire accounting process. Accounting documents like performa invoices, receipts, credit bills etc are processed automatically.
We will now understand the Straight Through Processing automation process and how exactly does it process documents without human inputs.
Straight Through Processing Automation Process
STP document processing is five crucial steps involved. We will deep dive into each step to understand how exactly the entire process is executed without any human intervention.
1.Data Capture Accuracy:
The first step once the documents are uploaded is to identify how accurately the data has been captured. Advanced OCR along with context based AI technology will detect, automatically extract and capture the fields. Invoice numbers, billing amount, date, invoice total, supplier name, customer name, currency, line item description, line item quantity, line item rate, line item code, line item total etc are some of the many fields that will be captured automatically.
Once the data is captured, it can be manually checked whether right information has been extracted for the right field. In case it has not, it can be amended and next time the similar document is uploaded, the self-learning AI technology will recognize it and capture the data accordingly.
After the data is captured by the system, a confidence score is generated. In very simple terms, the confidence score is a metric to understand the probability of accuracy of information from the documents. The confidence score is based on the fields that have captured data, the quality of the document scanned and so on.
KlearStack AI calculates confidence score per field and the overall confidence score for the document. KlearStack AI recommends to keep the minimum confidence score threshold at 85% or higher to extract and process documents as accurately as possible. However, organizations can amend the confidence score threshold as per their process tolerance. This option is given to the KlearStack users as each organization has unique processes to handle invoices, receipts and such documents.
So if the organization has set the confidence score to 70% and document uploaded shows a confidence score of 65, KlearStack STP will flag it as an exception and stop the document automation process. If the confidence score of a document uploaded is 75, documents will proceed to next stage of STP – Data Validation.
At this stage, customizable data validation rules set by the organization are matched with the documents uploaded. There is no limit as to how many data validation criteria can be set by the organization. This stage helps to filter out documents that should proceed with the next steps of STP.
Say the organization wants to have at least 1 manual review and approval step for invoices above USD 50,000 value. If this data validation rule is set by the organization, invoices that have total billing amount above USD 50,000, will be sent to the next stage. If the value of an invoice is less than USD 50,000, KlearStack Straight Through Processing will not allow the documents to proceed to the next stage.
This is the second last stage in the STP automation process. User can set cross document type reconciliation rules for checking data integrity across multiple data sources. E.g. the PO reference number from a newly processed invoice can be cross checked with PO number from corresponding purchase order. This process is very important because it validates if the transaction matches with previously processed document of the same number or not, that had the exact same item descriptions, quantity of goods and total billing amount.
Let us say the customer raised a purchase order with the number PO1234 towards a supplier. Once the order is fulfilled, the supplier will share the invoice with the customer with the same purchase order number, PO1234. The system at this stage will reconcile these documents. It will match the purchase order of the supplier and invoice of the manufacturer to ensure that documents of same transaction are being processed.
This is the final stage of STP document automation process. The original predicted data from the documents is transformed and converted as per the user’s needs. If for example the supplier is using alpha numeric for their stock inventory identification codes, and the customer is using alpha numeric codes, the documents will match and transform codes as per the suppliers’ numeric codes and customer’s alpha numeric code format.
With this, STP document process is over and the extracted data will be archived on the QuickBooks or SAP integrated software. If KlearStack API is used, the data can be then stored with any integration.
Why Choose KlearStack?
KlearStack AI provides template free data extraction. The contextual based AI technology with the help of image processing and computer vision tools, data can be captured, classified and stored from any document format. KlearStack AI API can be integrated easily with any of the existing solutions. If you would like to automate your document processing needs, click here to get in touch with us.