Tier 1 US Bank
This multi-currency FX trade to cash reconciliation was being performed using an in-house customer-developed machine learning reconciliation system. With user-supplied tuning, it was giving superior match rates when compared to their previous rules-based matching. Even with this effective customer machine learning, the reconciliations team still had to manually match thousands of transactions each day, making this tricky reconciliation in the words of the operations teams a 'nightmare’.
After installing Matchimus the automatic training was run once with a single customer extract of one week worth of data. Matchimus was then able to boost the customer’s matching rates from 93% to over 97%, delivering a 60% reduction in manual matching. A match accuracy of 99.99% achieved a 95% reduction in mismatches (which were previously occurring at a rate of 0.3%)
The customer operations strategy team called implementing
Matchimus a "no-brainer" and “I wish all our STP implementations had this guaranteed ROI”
This cash to cash reconciliation was being performed with a market leading rules engine using over 30 customer-crafted rules which were delivering a respectable auto match rate of close to 79%, but this still left thousands of transactions to be manually matched every day.
Using a single training run on historical customer matches and with the use of one customer defined grouping definition, Matchimus was immediately able to deliver a boosted match rate of 94%, yielding a more than 75% reduction in manual matching and a 100% reduction in mismatches (which were occurring at a rate of 0.5%)
This retailers' daily expense reconciliation was being performed with using a market leading cloud-based reconciliation engine using 13 customer-crafted rules. Despite a good match rate of 93% the users still had to manually match thousands of transactions and were planning to move the reconciliations team to a lower cost location.
Following a single training run using historic data, Matchimus was able to boost the customer matching rate from 93% to 98%, an 80% reduction in manual matching work and along with a 95% reduction in customer mismatches.