Verusen and AI-Powered MRO Optimization: Freeing Working Capital From Messy Materials Data
Walk into any large manufacturer and you will find millions of dollars sitting on shelves in the form of spare parts. Bearings, valves, motors, filters, the maintenance, repair, and operations (MRO) materials that keep the plant running. Individually cheap, collectively enormous, and almost always mismanaged. The same part is stocked five times under five different descriptions across five plants, some sit unused for years while others run out at the worst moment, and no one has a clean, single view of what the company actually owns. It is one of the largest pools of trapped working capital in industrial business, and it persists because the underlying data is a mess.
An Atlanta company called Verusen has built AI to fix exactly this. Its platform harmonizes disparate MRO data across a manufacturer's many systems, gives supply chain teams true visibility, and optimizes what to stock and where. This is a deep look at what Verusen does, the results it delivers, and, just as importantly, what it actually takes to get value from it, because with a product like this the value is gated almost entirely on one hard thing: the data.
The problem Verusen set out to solve
MRO materials management is a data problem masquerading as an inventory problem. A global manufacturer runs multiple ERP systems, often one per site or per acquisition, each with its own way of describing parts. The result is chaos: duplicate records, inconsistent descriptions, no way to see that the valve needed in one plant is sitting idle in another. Because no one can trust the data, everyone over-stocks to be safe, and working capital bleeds into the warehouse.
Verusen was founded in 2015 to attack that root cause. Under founder Paul Noble, now Chief Strategy Officer and a board director, and CEO Scott Matthews, the company built what it calls the Material Graph, described as the world's largest MRO materials knowledge base, having ingested more than 41 million unique SKUs and $12 billion in annual inventory and spend. The thesis is that you cannot optimize what you cannot see, so the first job is to use AI to harmonize the messy data into a single trustworthy view, and only then to optimize the inventory on top of it.
What Verusen actually does
Verusen is not a dashboard bolted onto an ERP. It is an AI layer that turns fragmented materials data into decisions. According to the Verusen solution, the platform works in a logical sequence:
- Harmonize the disparate MRO data across every enterprise system into a single, deduplicated, trustworthy source, using LLMs, data science, and the Material Graph.
- See true visibility of materials, inventory, and spend across all sites at once.
- Optimize stocking levels, procurement, and inventory planning based on that clean, connected picture.
The company has also moved to make the AI explainable, launching an explainability agent so supply chain teams can understand the reasoning behind a recommendation rather than trusting a black box. That matters in an operational setting where a planner is being asked to reduce a safety stock they have relied on for years. Verusen is now extending beyond indirect MRO into direct materials, components, and raw materials, across both discrete and process manufacturing, and into sectors like oil and gas, mining, and utilities.
Who benefits, and the proof in the numbers
The beneficiaries are concrete. Supply chain and procurement teams get a single trustworthy view instead of a pile of conflicting spreadsheets. Maintenance teams get the right part when they need it. And the CFO gets working capital back off the shelf.
The reported results are the kind finance notices. Customers typically reduce MRO inventory by 15 to 25 percent, unlocking millions in working capital. In one case, a Fortune 500 industrial manufacturer connected four ERPs through Verusen and cut MRO working capital by 16 percent in six months without a single stockout, which is the crucial detail: freeing cash without creating operational risk. Verusen has more than 30 customers, was named to the Inc. 5000 fastest-growing companies list in consecutive years, and secured $25 million in Series B funding as part of roughly $39 million raised to date, reported by outlets including Logistics Management.
Why this matters beyond MRO
Verusen is a strong example of the pattern defining enterprise AI's winners. They are not general-purpose assistants. They go deep into one high-value, data-heavy problem, build a proprietary knowledge asset that compounds, and earn trust by making their reasoning explainable. MRO optimization is a near-perfect target: enormous trapped value, messy inputs, and a clear financial payoff for every point of improvement.
It also points at the real bottleneck in enterprise AI, which is not the model. Verusen's own product order tells the story: harmonize first, optimize second. The optimization is the easy, valuable part. The harmonization, cleaning and connecting years of messy data across many systems, is the hard part, and it is where the entire outcome is won or lost. This is the same lesson showing up across every category: the intelligence is ready, but turning it into a result inside a real enterprise is the work. It is why forward-looking manufacturers are rethinking how they put delivery capacity to work, toward a new era of execution built on delivered outcomes rather than rented effort.
Connecting the software is easy. Cleaning the data is the project
Verusen can be connected to a manufacturer's systems, and the platform does the heavy lifting of harmonization with AI. But getting to trustworthy, optimized inventory across a global manufacturer is a project, and it has several distinct pieces, none of which happen on their own.
1. Data harmonization and master-data cleanup. This is the heart of it. Even with powerful AI doing the matching, getting a manufacturer's materials data genuinely clean, deduplicated, and mapped into a single taxonomy takes real work: validating matches, resolving edge cases, and building the master-data governance so it stays clean. The quality of everything downstream depends on this, and it is where value is created or lost.
2. Multi-system and ERP integration. The value only appears when all the sources are connected. Wiring Verusen into multiple ERPs, plus procurement systems and maintenance or asset-management platforms (EAM/CMMS), reliably and across sites, is substantial integration work, and the more plants and acquisitions a company has, the bigger it gets.
3. Taxonomy, criticality, and policy configuration. Optimization is only trustworthy when it reflects the company's reality: which parts are critical, what the real lead times are, what stocking policies apply by site and by material class. Encoding that, and tuning the recommendations to the plant's actual risk tolerance, is expert work that determines whether planners act on the output.
4. Change management with planners and procurement. This is the piece that decides success. You are asking experienced planners and maintenance leaders to reduce safety stocks they have trusted for years. If they do not believe the recommendation, they will ignore it and the working-capital gain never materializes. Building that trust takes the explainability, a rollout sequenced to visible wins, and workflows that put the AI's insight where planners already work.
5. Measurement and continuous tuning. The value case has to be proven in the manufacturer's own numbers. Baseline inventory value, working capital, stockout rates, and turns before, measure the same after, plant by plant, and keep tuning. Without that loop, you cannot prove the savings to finance or justify rolling it out to the next site.
Why manufacturers stall, and it is not the software
None of those five require a better model. The bottleneck is capacity and specialized know-how, not capability. Getting Verusen fully live across a global manufacturer is a cross-functional effort spanning supply chain, IT, data, and plant operations, and most manufacturers do not have a spare team that understands both materials data and AI deployment sitting idle to run it. A lean, fast-growing vendor keeps its own people focused on the product and the largest accounts. The manufacturer's planners are busy running the plant. So the rollout stalls at one or two sites, the data cleanup drags, and the enterprise-wide working-capital release never fully arrives.
This is not a Verusen problem or a manufacturing problem. It is the defining problem of enterprise AI. Hiring a permanent data-and-deployment team for a one-time rollout is slow and expensive, and traditional staffing rents bodies by the month rather than delivering a working outcome.
Closing the gap faster
Manufacturers that win with supply chain AI treat deployment as the real project and resource it properly. That usually means a delivery layer that sits behind the vendor's product and the manufacturer's team and owns the last mile: driving the data harmonization and master-data governance, integrating the ERPs and operational systems, configuring the taxonomy and policies, running the change management with planners, and building the measurement loop that proves the savings.
This is exactly what a Virtual Delivery Center is built to do. AiDOOS runs elastic delivery pods, guided by a proven delivery playbook, that plug in alongside a product like Verusen and the manufacturer's own people, take on the data and deployment work, and get the platform live and releasing working capital site by site. The vendor keeps its focus on the product. The manufacturer reaches results without hiring a permanent function for a one-time push, with the accountability of a managed delivery model and payment for delivered outcomes rather than effort.
The sensible way to start is small and provable: pick one plant or one material category, get the data harmonized and the optimization trusted and live, measure the working capital freed, and use that proof to scale across the enterprise.
The bottom line
Verusen has taken one of manufacturing's largest pools of trapped capital and built AI that turns messy materials data into visibility and savings. The consistent 15-to-25 percent inventory reductions, the Fortune 500 results, and the Series B backing say the approach is real. The technology is ready.
The question every manufacturer should ask is not whether the AI can optimize inventory. It is whether the company will do the data work, connect the systems, earn the planners' trust, and prove the savings, or whether the initiative will stall at one plant with the data half-cleaned. Answer that well, and millions in working capital come off the shelf. Leave it to chance, and even a strong platform like Verusen becomes another project that started well and never scaled. The model is no longer the hard part. The data and the deployment are, and that is a solvable problem.
If you are deploying Verusen or any data-heavy enterprise AI, AiDOOS provides the delivery pods that get it live and proving ROI. To explore the product itself, visit Verusen. To see how enterprises put managed delivery to work, explore how enterprises use AiDOOS.