Logistics: Escape from the Jungle of Suspicious Packages

​A company is as effective as its least stable system or process. Our Client, who is in the business of procuring and selling small medical equipment, has experienced it at first hand. When they contacted us, all but one warehouse could meet an abundance of orders with ease, partly because they had a stable software system in place covering warehouse management, multicarrier shipping, label printing and other logistics and business operations. Meanwhile, the warehouse managing export orders could barely handle the workload despite the extra hours and intense work pace, even though they were armed with the same software solutions and high-quality staff as the rest of the warehouses.

It turned out that this uniformity of systems and processes was exactly the problem. They were managing their operations with the very same, well-built and stable software that was used in all locations while their workflow contained an important additional phase. This industry is fraught with regulatory obligations, as some of the equipment sold by the company had serious potential for misuse (for example, a small centrifuge that is used by a diagnostic lab for blood fractionation may be used to create illegal substances including explosives, when in the wrong hands). Thus, each export shipment should be screened and filtered to cancel any suspicious transactions including blacklisted companies, tax fraud in form of constant return of equipment, and especially companies who pass on equipment to unauthorized parties or sanctioned countries.

The customer had access to several domestic and international official sanction lists and third-party due-diligence databases for the screening process. For instance, there were lists from governmental treasury departments with entities and individuals that require additional screening or were summarily prohibited from receiving certain technological implements. Examples of these lists include BIS Denied Persons List, Unverified Entities List, EAR-Entity List, NPT sanctions, OFAC SDN and FinCEN lists, Australian DFAT Consolidated List, Canadian OSFI Cumulative Warning List, HM Treasury Sanction List, PEP lists, and in some cases Lloyd’s List.
​
Employees received the list of the daily shipments a day prior shipment. They mechanically checked each recipient and ultimate consignee in every relevant sanction list before forwarding it to label printing and dispatching. This mechanical check approximately translates to 6 hours of work per employee, while only a small portion of the orders, not more than 10% was actually deemed suspicious. Besides, searching separately in each list did not enable the discovery of cross-list relations, which threatened compliance. With the existing workload, shipping deadlines were constantly slipping. The process also impacted employee morale, as due-diligence tasks weren’t part of the official duties for most of them. This was a conundrum, as hiring a multitude of dedicated due-diligence personnel was unviable for the business. Errors, slips, post-facto regulatory enquiries, and ebbing morale were compounding an already-difficult problem.

Our mission was to develop a custom software that seamlessly integrated with the existing systems, automated screening, armed human experts with crucial alerts and flags, but left the critical decisions to human expertise through a simple UI. Our solution was based on a seamless integration of a several sanction lists (custom lists as well as published lists) for due diligence. First, we examined the structure of the database and created an adapter to each list. The next step was a module that compared each recipient and ultimate consignee with the sanction lists. Each order received a match score (where, for example, a mere first or surname match received a low score, while a company name – address match or company name – recipient match was given a higher score). Certain sanction lists had a much higher score than the others. The scoring system was created based on the Client’s expertise and experience.

In addition to such automated validation against a consolidated screening list, a system of flags and alerts based on simple rules and heuristics was also incorporated into the scoring system. For instance, red flags were created if the recipient’s address was residential, the company wasn’t recognized as a medical or pharmaceutical lab, payment was in cash instead of international bank transfer, the payment originated from a third country not involved in the consignment, the ultimate consignee was a “remote” location, there were incomplete entries in forms relating to SED, or there were significantly more purchases from a location or recipient than the average in a given period of time, etc.

Orders below a preset score threshold were categorized as non-suspicious and cleared for shipping. For the rest that crossed the threshold, emails and push notifications were sent to employees designated as internal compliance officers. These designated employees accessed the suspicious orders on a user interface where the risk levels were indicated with color codes. Experts had three options for each suspicious order: they could waive the alert and clear for shipment, block the shipment and initiate related steps, or suspend the shipment for further investigation. Advanced search options and an email system with tracked conversations were among the tools available for further investigation and formalities. With this system, label printing and dispatching happened in parallel with the vetting process and remedial steps, eliminating workflow bottlenecks. Blocked orders and suspended orders were listed in batches of simple-but-comprehensive CSV files and were sent to another third-party software module in the customer’s ecosystem to perform administrative and reporting tasks.
​
Scoring system and risk levels were finetuned during the test period based on employee feedback. Since the final custom software solution went live, the throughput at this warehouse has increased by 62%. In part, this is thanks to the fact that this new system enables 90% of the consignments that are beyond suspicion to go through without manual verification. In other words, manual due-diligence is now focused on about 500 shipments a week, out of a total of (say) 5000 shipments. The warehouse can now manage the daily orders seamlessly and utilize the full capacity of other company software systems.