IHS Markit, a firm that has won more awards than any other across the four brands nestling under the WatersTechnology umbrella, wins yet another category, thanks to its Commodity Tracker, which beat five other entries for the win. The offering, which went live at a US-headquartered tier-one investment bank in May last year, automates the largely manual processes around how commodities traders track inventory and supply. There has traditionally been a disconnect between trading, which is increasingly electronic, and the physical world of inventory management. Commodity trading firms typically require large numbers of warehouses and transporters, each reporting information in different formats about shipments. Accurately managing inventory is therefore manually intensive, complex and operationally risky, and involves time-consuming reconciliations and high fixed costs.
The platform digitizes the tracking of base-metals inventory for market participants, and includes a number of innovative technologies such as artificial intelligence (AI) alongside more mature technology applications. The heart of Commodity Tracker features optical character recognition (OCR) technology that reads image files such as PDFs and JPEGs, and creates digital characters of the data within those files. While OCR is a mature technology, its effectiveness is contingent on the image presented to it, often resulting in imperfect data being produced. This shortcoming is addressed by way of machine-learning (ML) technology, which organizes the digital data, based on the training created. The data is also automatically normalized to standardize areas such as weight units and date formats. With this approach, the more documents Commodity Tracker consumes, the more accurate it becomes. The platform, hosted by Amazon Web Services, was designed with a micro-services architecture allowing flexibility on feature development, while the use of Red Hat’s OpenShift platform employs container management and orchestration development and onward support.
According to IHS Markit, the benefits realized by the bank on the back of Commodity Tracker include: replacing the manual reconciliation of 3,500 documents each month from 150 metals warehouses; a 50 percent time reduction in performing monthly reconciliations; faster discovery and resolution of breaks during the reconciliation process; and more transparency about trends in data which helps identify risk
Recently, the bank was able to use Commodity Tracker’s inventory-mapping feature to visualize and review its metal stockpiles in the Gulf Coast area, in preparation for Hurricane Harvey. By reviewing the predicted path of the hurricane against the bank’s mapped inventory, the team was able to notify insurers, plan for shipment delays, and create contingency plans where there was risk to contractual obligations in terms of delivering material.
Bryan Cross, who heads UBS Asset Management's QED group, joins to discuss alternative data and AI.Subscribe to Weekly Wrap emails