June 16, 2016 | Spend Analysis
Uber recently announced their next big idea: Driverless cars taking over the streets! Let’s be honest, the thought of sitting in a driverless car bringing you home from work in the evening, may initially give you the shivers. But, we have already experienced automation in almost every industry. We deposit our cash and checks into ATMs, we speak to machines when booking a hotel room, we manufacture cars using industrial robots - why could they not drive? A more difficult question would be: Will we ever see complete automation in Spend Analysis?
For years Supply Chain Management Technology companies have been striving to achieve automation in spend analysis. Today we use Artificial Intelligence with great results and without human interaction; but, they somehow struggle when applying the Artificial Intelligence to the data received overall, as data itself includes errors.
Spend Analysis starts at the transaction entry level by the client. When a purchase order or an invoice is entered into the system, sourcing team members are requested to enter various pieces of information such as material descriptions, GL account number, business group, region and others that are used for the Spend Analysis. In our experience, we have seen so many missing or inaccurate data points that we eventually had to manually review the data and come up with ways to identify how the data can be categorized.
How can we ensure the data received from clients’ ERP systems are clean and accurate? The quick answer would be to train them. Sourcing team members need to be aware of Spend Analysis methods, and the need for accurate data entry. Another answer is to enable ERP technologies that will require complete and accurate data entry.
So, will we ever achieve complete automation in Spend Analysis? Apart from state of the art technologies and the AI Supply Chain Management Technology Companies use, the answer depends on the accuracy and quality of data clients keep in their ERP systems. Once perfection is achieved there, we can also achieve “driverless” Spend Analysis.