When an event arises that does not fit the pre-defined smart contract parameters it generates an exception. This could be caused by numerous circumstances, including a variance in quantity shipped versus quantity ordered, a price discrepancy, or an unexpected adjustment.  These exceptions though are often repeated under similar circumstances, or perhaps with a specific trading partner.  Using artificial intelligence and machine learning techniques, patterns are recognized and learned to enable even complex exceptions to be automated where they fit a known pattern with a consistent expected response. 

Handling these exceptions more quickly and without human intervention avoids delays and accelerates transactions. Also, the  automated  handling of these exceptions without having to refer back to the trading partner improves the quality and timeliness of trading activities.