The pattern Kaunt finds in your accounting processes is determined by many different elements. Data from your historical processes are used, and thus, large amounts of data is involved in finding a pattern.
Kaunt scans the invoice and, depending on the available information, the robot takes up to 30 different elements into account as it is being trained and is finding a pattern. If one data element is dominating, e.g., supplier, area code, VAT number or such, this scores higher for relevance. However, this is not enough to create a pattern, and more data points will always be taken into account.
From the many different data points that are scanned by the robot, the deciding factor differs depending on the invoice and the historical data. There will be dominating deciding factors, which together with all other data points form the foundation for the patterns that are replicable.
Besides data on the invoices, the other important element is your accounting processes, history and your consistency. It is important that you are consistent in your accounting processes to provide Kaunt with a strong foundation for the best possible conditions to find the right pattern between the invoices and your accounts. You can read more about what makes good accounting processes here.
Kaunt can process an invoice down to the different lines and add the specific line to the right cost centre. Again consistency is the key and you need to be consistent in your booked invoices, for kaunt to do so.
Due to the different data points that are part of the pattern recognition, the robot has a strong foundation for booking invoices correctly. Hence, the many data points give the robot a security when booking invoices and allows the robot to be selective and employ the relevant data, to ensure that the booking occurs on the correct account.
Whether Kaunt books or registers for an employee to approve, the robot operates with an identical level of security in both scenarios. The robot is always taking new data into account and is trained all the time. So, if the robot is not sure how to process the data, the invoices are not processed by the robot, but will be handled manually by an employee.