Trending
Written by: Jerry Glynn
For three centuries, business strategy came down to three levers: Land. Labor. Capital. Economists taught us that whoever controlled the scarce resources controlled the market. That model defined the industrial world, and many of the leaders I work with adhere to it.
Economist Edward Yardeni has identified what he calls the “fourth factor of production” — data.1 Unlike the first three, data is not scarce. It does not deplete. It does not require land or machinery to produce. Every mid-market company that is sitting on years of transactional history, customer behavior, operational data, and supply chain records is already holding an unlimited — and invaluable — resource.
The problem has never been the data. The problem has been the cost and complexity of unlocking it. Until now.
The Legacy Tax Is Eating You Alive
As a senior technology executive who has worked alongside CEOs, and operational leadership teams at mid-market companies, I have seen this pattern consistently: the data exists, but getting to it reliably is a major — and expensive — undertaking. Teams spend months trying to reconcile reports from systems that do not talk to each other. Analysts rebuild the same spreadsheets every quarter. Decisions get made on gut instinct because the data pipeline is either broken, delayed, or simply inaccessible.
A recent Gartner study confirms that this is not a perception problem. More than half of supply chain leaders — 56 percent — cite integrating AI with legacy systems as a major challenge.2 The technical debt is real. But so is the breakthrough approach now available to address it.
AI Is the Refinery, Not the Replacement
Here is the strategic shift every mid-market CEO must consider: you do not have to fix the plumbing to use the water.
Modern AI acts as an intelligence layer that sits above your existing systems. It gleans insights and foresights across your disconnected data sources, translating the “Tower of Babel” systems rife with raw, chaotic input into clear, actionable output. Your legacy systems become the raw material. The AI becomes the refinery. You stop paying to maintain the status quo and start extracting value from what you already own.
The evidence that this works is no longer theoretical.
Ascendum is a global machinery and equipment distributor operating in 14 countries. Their field service teams were drowning in disconnected data: customer records in one system, technical specifications in another, manufacturer documentation scattered across more than 14,000 PDF files. Diagnosing an equipment issue in the field took up to 30 minutes of manual searching across multiple databases. Rather than rebuilding those systems, Ascendum layered a generative AI solution over the existing data infrastructure. The result: troubleshooting time dropped from 30 minutes to under a minute, and first-contact resolution rates improved by 50 percent.3
SpecialKids, a specialty retailer of adaptive clothing for children and adults with special needs, faced a familiar scaling problem. As the business grew more than 40 percent year over year, inventory planning across hundreds of SKUs and multiple sales channels became unmanageable on spreadsheets and a legacy ERP that could not reconcile demand signals reliably. They deployed Inventory Planner by Sage (a leading SMB business management platform) — an AI-driven forecasting layer — directly on top of their existing operational data, without an ERP rip-and-replace. After deploying AI-informed forecasting, stockouts fell by 77 percent. Overstock dropped by a third, freeing up capital tied to the wrong SKUs, and the inventory team reclaimed more than 20 hours a month previously lost to manual spreadsheet work.4
Nolinor Aviation, a Canadian charter airline and one of the world’s largest operators of the Boeing 737-200, faced a legacy data problem with real safety stakes. Nolinor Aviation’s safety investigations had been a grueling manual process: investigators parsing free-form safety reports, unstructured risk documentation, and years of compliance records by hand. Each complex case took up to 40 hours of human effort. Working with AI research institute Mila and technology partner P3F, Nolinor built an “assistant investigator” — an AI layer designed not to replace the investigator, but to work alongside one. The system automates report structuring, barrier analysis, and technical documentation review, freeing the human expert to focus on interpretation, judgment, and accountability. The result: complex investigations that took 40 hours now require approximately five hours of human involvement — an 80 percent reduction.5
Where to Start: Three Moves in 90 Days
You do not need a multiyear roadmap to begin. The companies forging ahead are making targeted moves. My advice:
· First, take inventory before you take action — and make it a business exercise, not an IT project. Map where your most critical data actually lives: which systems hold it, who owns it, how current it is, and where the integration gaps are. Critically, an operational leader — not just your IT department — must own the data definitions and the intended business outcomes. IT can tell you where the data lives. Only your business leaders can tell you what it should mean. This data readiness assessment does more than identify problems; it reveals your data maturity and tells you exactly which high-value use case to initially pursue.
· Second, choose one high-value, high-pain use case. Not a pilot. A business problem with a measurable outcome — inventory accuracy, service resolution time, working capital, customer churn. Deploy an AI intelligence layer targeting that single problem using your existing data infrastructure. Prove the model, then scale it.7
· Third, evolve the stack, do not replace it. Gartner’s guidance to supply chain leaders applies broadly: deliberately evolve your technology rather than attempting wholesale replacement.8 Build the AI and data layer above your data. Swap components as you prove value. Protect the operation while you build the advantage.
The Window Is Open. It Will Not Stay That Way.
Mid-market companies that will lead throughout the next decade are not waiting for the perfect technology environment. They are treating their existing data as the strategic asset Yardeni described: unlimited, already owned, and ready to be refined.
AI has lowered the cost of data integration to a point where size is no longer a barrier to extracting value from existing data. Capabilities that were once available only to the largest, most heavily resourced enterprises are now accessible to any company willing to effectively pull the data lever.
Related: Should SpaceX Have a Place in a Core-Satellite Portfolio?
Citations
1 Ed Yardeni’s framing of data as a fourth factor of production comes from Yardeni Research’s May 2026 commentary, where he argues that land, labor, and capital are scarce, while data is unlimited but historically expensive to collect, process, and analyze..
2 Gartner, "Gartner Survey Finds Technology Integration and Talent Perceived as Key Roadblocks to Scaling AI in Supply Chain," April 29, 2026. Survey of 140 senior supply chain leaders from organizations with annual revenues of $250 million or more.
3 McKinsey & Company, "From Pilot to Profit: Scaling Gen AI in Aftermarket and Field Services," March 2025; and "Adding a Powerful New Tool to the Field Mechanic's Toolbelt: AI," Case Study, 2024. Ascendum is a Portugal-based global provider of machinery and equipment with operations in 14 countries and annual turnover of €1.3 billion. The new system increased first-contact resolution rates by 50 percent and reduced troubleshooting time from 30 minutes to under a minute.
4 Inventory Planner by Sage, "SpecialKids Slashes Stockouts by 77% with Inventory Planner's Perfect-fit Solution," Customer Story, inventory-planner.com. Results include a 77% reduction in stockouts, a one-third reduction in overstock, and 20+ hours per month saved on manual inventory processes.
5 The AI Journal, "P3F at the G7 Summit with an AI Innovation in Aviation Safety," June 10, 2025. Nolinor Aviation partnered with AI research institute Mila and technology firm P3F to develop an AI assistant investigator for safety compliance workflows. The solution reduced complex investigations from 40 hours to approximately 5 hours of human involvement — an 80% reduction — while keeping the human investigator responsible for all final conclusions.
6 Salesforce, "New Research Reveals SMBs with AI Adoption See Stronger Revenue Growth," December 2024. Survey of 3,350 SMB leaders globally. Growing SMBs are twice as likely to have an integrated tech stack (66% vs. 32%) compared to SMBs with declining revenue.
7 McKinsey & Company, "Building the Foundations for Agentic AI at Scale," April 2026. Rather than rebuilding systems from scratch, companies can modernize each layer to improve visibility and governance, deploying an AI layer that decouples capabilities from the underlying legacy infrastructure.
8 Gartner, "Gartner Survey Finds Technology Integration and Talent Perceived as Key Roadblocks to Scaling AI in Supply Chain," April 2026. Gartner advises leaders to "deliberately evolve—rather than replace—their tech stacks, building a unified data layer and agentic layer that sits atop legacy systems."