Closing the gap between contract and execution

Health systems negotiate significant savings through contracts, standardization initiatives and strategic sourcing programs. Yet some of that value is lost before it reaches the organization’s profit-and-loss statement.

Pricing discrepancies, off-contract purchases, unmanaged spend, and inconsistent utilization can create gaps between the savings negotiated and the savings realized. Fragmented technology systems make the problem more difficult because contracting, purchasing, clinical, and financial data frequently reside in separate platforms.

Selles said tens of billions of dollars in potential non-labor value are lost between contracting and execution across the healthcare industry.

“The real opportunity sits in the many steps between contract, purchase, payment and utilization, where value is either captured or lost,” he said.

That challenge is changing the role of group purchasing organizations. Rather than focusing primarily on negotiating prices, GPOs are becoming more involved in helping health systems manage what happens after contracts are signed.

Vizient, for example, is expanding capabilities such as price assurance and procure-to-pay integration to reduce discrepancies between contracts and invoices. Tail-spend automation can bring smaller, unmanaged purchases under greater control, while clinically integrated supply chain programs connect product decisions with costs and patient outcomes.

“This is not about adding more services around the edges,” Selles said. “It is about being embedded in the workflows where purchasing and utilization decisions happen, so value is protected at the point of execution.”

The shift reflects the financial pressures confronting health systems. According to Selles, non-labor costs have risen 50% since 2019. As a result, chief financial officers are asking supply chain leaders for more than retrospective reports or additional dashboards.

They want price accuracy across contracts, purchase orders, and invoices; real-time insight into high-cost clinical utilization; and stronger contract compliance across hospitals and ambulatory settings. Most importantly, they want tools incorporated into existing workflows rather than added as separate systems employees must remember to use.

“What health systems are asking for now is intervention, not just information,” Selles said.

Moving from reactive to predictive

The same distinction is shaping how health systems manage supply disruptions.

Historically, many organizations relied on transaction data, manual processes and disconnected systems to identify shortages. Supply chain teams often learned about a disruption only after orders had been delayed or products were placed on back order.

Brian Wells, senior vice president of sales operations at Medline, said the industry is beginning to replace this reactive approach with predictive, intelligence-driven operations.

“We’re seeing a shift toward predictive models that use real-time data, AI, and advanced analytics to identify risks before they disrupt operations,” Wells said.

Instead of notifying a health system that a shortage is already underway, predictive technology can help determine which products are at risk, when a disruption may occur, how long it could last, and which alternatives are available.

Medline has applied that concept through Mpower, an AI-powered digital control tower that brings together the company’s manufacturing, distribution, inventory, and fulfillment data. The platform analyzes supply information to identify at-risk stock-keeping units, estimate potential disruption and recovery timelines, and recommend substitute products.

That end-to-end view can provide information that may not be available through an individual health system’s ERP platform, which typically reflects only part of the product journey.

“Mpower doesn’t simply report what’s happening—it helps drive action by connecting supply chain, value analysis, and clinical teams through automated workflows, substitution recommendations, and coordinated communications,” Wells said.

That coordination is critical because responding to a disruption often requires input from several departments. Supply chain may identify the shortage, but clinicians and value analysis teams must evaluate whether an alternative is appropriate. Purchasing and operational teams must then implement the change and communicate it across affected locations.

Predictive intelligence gives those stakeholders more time to evaluate alternatives and complete approvals before product availability affects care delivery.

Measuring the operational impact

The value of intelligent supply chain technology must ultimately be demonstrated through measurable operational and financial results.

During an initial Mpower pilot, risk-resolution cycle times decreased by more than 50%, falling from 5-7 business days to approximately 2.5 days. More than 300 at-risk products were addressed through early substitution approvals during the first 30 days, according to Medline. Pilot customers using the company’s pre-approved AutoSub program also recorded a 1-2% increase in unadjusted fill rates.

Wells said health systems should assess these tools using a combination of efficiency and resiliency metrics, including:

• Time required to identify and mitigate a supply risk.
• Number of disruptions addressed before operations are affected.
• Fill-rate improvement.
• Reductions in urgent escalations.
• Consistency of substitution and approval workflows.
• Staff time redirected from crisis management to strategic work.

Other technologies are delivering results in different parts of supply chain operations. Selles said warehouse automation currently offers some of the clearest ROI because its labor savings are measurable and established, even though warehousing accounts for only part of total supply chain spending.

AI-supported forecasting is improving inventory management, fill rates and operational consistency. Real-time location systems can reduce the time clinical teams spend searching for equipment, while RFID technology continues to expand within healthcare inventory and asset management.

The larger opportunity, however, lies in connecting these tools.

That opportunity is also driving larger technology investments across the GPO sector. In February, Premier announced a multi-year program to modernize its data infrastructure and develop an AI-powered decision intelligence platform intended to connect data, analytics and execution across healthcare supply chain operations. The company is working with Palantir Technologies, Databricks, and Microsoft to bring operational AI, unified data capabilities, and cloud infrastructure into the platform.

Premier said the initial work will focus on real-time visibility into non-labor spending, identifying savings at the point of order, detecting supply risks before they affect care, and aligning contract prices across vendors. The initiative illustrates how the intelligent supply chain is evolving from a collection of individual technologies into a broader intelligence layer spanning supply chain, financial, and clinical operations.

ERP platforms can provide foundational operational data. Automation can eliminate repetitive work. AI can analyze patterns and identify emerging risks. When implemented separately, each may deliver incremental improvements. When connected through common workflows, they can support decisions throughout the supply chain lifecycle.

In one pilot with a large Midwestern health system, Medline combined inventory review, process standardization, storage redesign, predictive analytics, automated substitution, AI-enabled inventory visibility, electronic shelf labels, PAR optimization, and emergency preparedness planning.

The effort reinforced that technology alone does not create an intelligent supply chain.

“Meaningful transformation can happen when people, processes and technology are aligned around a shared operating model,” Wells said.

Solving the data disconnect

The potential of AI and predictive analytics still depends on the quality and connectivity of the underlying data.

Many health systems have large volumes of contracting, purchasing, inventory, clinical, and financial data. The problem is that those data sets frequently do not communicate with one another or arrive too late to influence a decision.

“The biggest challenge is not access to data; it is that the data is disconnected,” Selles said.

Retrospective data can show leaders where spending exceeded expectations or where a product decision failed to produce the projected savings. But intelligent supply chain systems must surface that information while purchasing and utilization decisions are being made.

Adding more data without addressing integration can also create additional noise. Leaders may receive more reports and alerts without gaining a clear understanding of what action to take.

Premier’s technology roadmap is similarly built around interoperability, data governance, and the responsible use of AI in areas with clear operational impact. Databricks will support a unified environment spanning the company’s supply chain, clinical, and financial data, while Palantir will provide operational AI and analytics designed to support real-time decision-making. Microsoft will supply the cloud infrastructure for securely operating those workloads at scale.

The approach reflects a broader recognition that predictive analytics cannot deliver its full value while key information remains isolated. Premier plans to connect data and execution through a single intelligence layer, with the goal of supporting faster decisions, stronger margins, and more resilient operations.

AI can help organize that information, identify relevant patterns, and deliver recommendations at the point of decision. Vizient is working to embed AI into contract lifecycle management, procure-to-pay processes, and utilization workflows so insights appear within the work rather than through separate reporting tools.

Selles stressed that AI should strengthen human expertise, not replace it. Its role is to provide supply chain, clinical, and sourcing professionals with better information when they can still act on it.

That also requires closer alignment among departments that traditionally view performance through different lenses. Finance monitors cost, clinicians focus on outcomes and quality, and supply chain manages spending and availability. A clinically integrated supply chain brings those perspectives together, so product decisions reflect their combined impact.

“The opportunity ahead is not just better integration, but better orchestration,” Selles said.

Building resiliency without stockpiling

Supply chain intelligence is also changing the economics of resiliency.

Following the pandemic, many healthcare organizations increased inventory or established additional reserves to protect against shortages. Although additional stock can provide a buffer, it also ties up capital, requires storage space, and increases the possibility of product expiration.

Health systems are now looking for models that improve preparedness without relying exclusively on stockpiling.

“Resiliency is no longer a tradeoff against cost; it has to be designed into the supply chain,” Selles said.

Vizient Reserve, for example, gives participating organizations access to dedicated inventory without requiring each health system to hold all that product locally. Vizient said the program has helped create more than 180 million doses of essential medications and more than 3 million units of critical medical supplies through inventory commitments. More than 400 health systems have accessed over 5.8 million units of essential products through the program.

The model has been used during disruptions ranging from demand spikes of 280% during the COVID-19 pandemic to supply shocks affecting 60% of U.S. IV fluid production.

Predictive platforms complement these reserve models by helping organizations recognize threats earlier. Together, dedicated inventory, real-time visibility, and coordinated mitigation allow health systems to build readiness into their operating models rather than reconstructing a response during each shortage.

Designing AI around the people who use it

Even the most sophisticated platform will fail if employees cannot incorporate it into their daily work.

Medline collaborated with 10 U.S. health systems during Mpower’s development to understand how supply chain, clinical, and operational teams communicate, organize work, and make product decisions. One recurring challenge was not simply finding information but coordinating the right stakeholders around it.

That led Medline to design the platform as a shared source of information and collaboration, combining recommendations with automated approval, substitution, and communication workflows.

“Successful AI must reflect how organizations actually operate,” Wells said.

For health systems evaluating AI platforms, adoption should therefore be considered from the beginning. Leaders should examine whether a technology fits existing workflows, clarifies responsibilities, and reduces manual effort. They should also determine whether the platform helps teams act or merely produces another alert for employees to investigate.

The most useful systems will make the appropriate next step clear, bring affected stakeholders into the process and document how a decision was reached.

From purchasing power to performance

Over the next three to five years, intelligent sourcing and logistics capabilities are expected to become more embedded in daily operations.

Selles anticipates growth in agentic sourcing, which could allow health systems to address the long tail of lower-dollar suppliers and transactions that often remain unmanaged. Price-assurance capabilities will help verify that the contracted price is the price paid, while supply chain orchestration will connect planning, sourcing, and execution more closely.

Physical automation will also expand. Robotics and warehouse technologies could reduce operating costs and relieve labor pressures as health systems and distributors contend with ongoing workforce constraints.

Wells expects AI to play a broader role in supply chain analytics, inventory management, product standardization, and operational improvement. Across those applications, the objective will remain consistent: reduce complexity, coordinate teams, and help healthcare providers devote more time to patient care.

The intelligent supply chain, therefore, will not be defined by one platform, algorithm, or automation project. It will be defined by whether organizations can connect their technologies and expertise to produce timely, repeatable action.

For GPOs and health systems, that means shifting the focus from how much value is negotiated to how much is realized. For manufacturers and distributors, it means using end-to-end data to identify risks sooner and coordinate responses before care is disrupted.

The industry has spent years building visibility. Its next challenge is making that visibility operational.