Verusen: Revolutionary AI for the Supply Chain
Supply chain disruptions are a constant headache for manufacturers trying to keep production lines up and running. Shortages of material inputs and maintenance supplies can interrupt even the most streamlined processes. Those cargo ships lined up at ports around the globe underscore how hard it is right now to keep production lines up and running.
Today we’re announcing our lead investment in Verusen, whose Machine Learning software sits atop traditional ERPs to help manufacturers reduce the financial and operational burden associated with trusting their material supplies. Scale has been investing in supply chain technology since 2016 via investments in Airspace, Narvar, Locus Robotics, and KeepTruckin.
The problem Verusen is solving is very familiar to anyone with a manufacturing background or who has looked into the many causes of the supply chain breakdowns so often in the news these days.
Inventory Challenges Are Universal In Manufacturing
Getting inventory levels “just right” has two dimensions: financial and operational.
On the financial side (typically the purview of corporate executives), the focus is on avoiding inefficiencies with working capital. Too many dollars tied up in unneeded inventory is wasteful and adds up. These executives are tasked with reducing waste and ensuring working capital remains unrestrained.
On the operational side (typically the purview of plant-level managers and staff), the focus is on managing risk or quickly fixing supply shortages that impact the core productivity KPIs that they’re responsible for. From this perspective, excess inventory isn’t really a problem; it’s the time it takes to fix inventory problems that’s the problem because it takes time away from hitting production KPIs.
Those two sets of incentives are often at odds, as we heard in our conversations with Verusen customers. The typical materials management process reflects this. Inventory data lives in the corporate ERP but is inputted and maintained by plant-level staff via manual data entry. Company standards theoretically ensure uniformity between sites, but in practice compliance is often inconsistent and untimely. In the end, ERPs are full of duplicate entries and errors, resulting in frequent inefficiencies that prevent companies from knowing and trusting their materials data. Traditional data cleansing methods are tedious, expensive, and slow – and by the time you’re done, the “clean” data is already out of date. This is where Verusen comes in.
Verusen Introduces Materials Intelligence
Inventory optimization and the difficult orchestration needed to get it right is an almost universal pain point at manufacturers. It took Verusen’s revolutionary machine learning to change the dynamic. Big picture, the software is an intelligent layer sitting on top of the ERP platform, doing three big things:
- Replacing traditional, heavily manual cleansing of ERP records with self-learning, self-cleaning, always-on materials management intelligence;
- Harmonizing entries between different manufacturing sites, for example eliminating duplicate/substitute entries for the same material;
- Search enablement so that natural language searches can be used within material records.
Verusen’s automated cleansing functionality lays the groundwork for outsized impact from the other two optimization functions (better understood data allows better harmonization and search). The Verusen matching AI spans 80k different training points and is well tuned to inventory in particular. It is designed as a human-in-the-loop system in that the end user accepts/rejects Verusen recommendations for changes to materials data. Our research suggests that new customers experience a 15-20% acceptance rate of suggestions, climbing in the first year to about 50%, with some of the company’s early customers approaching 100% acceptance. This quick progression means its SaaS customers experience better performance every time they use the system.
Verusen is particularly valuable during this time of supply chain disruption. For the first time, different manufacturing sites inside a company can “share” inventory as needed. For example, a Verusen customer can locate a particular replacement material at a sister site and avoid, because of order backlogs, downtime spent waiting for that replacement part to arrive. This sort of inventory intelligence is truly revolutionary for many manufacturers.
A New Era in Supply Chain
We’ve been impressed with Verusen and its founder and CEO Paul Noble, since our first meeting back in 2019. Paul moved up through the ranks at The Sherwin-Williams Company where he specialized in supply chain/manufacturing and led its Eastern U.S. Industrial Distribution business unit. His extensive experience in the industrial supply chain and distribution space gives him the expertise to lead Verusen to new heights.
We look forward to working with Paul, the Verusen team, and the company’s investors in the years ahead.
Scale Supply Chain Investments
Originally published January 25, 2022.