Healthcare wants an autonomous supply chain but first, it has to connect one
AI-driven supply chains depend on trusted data, coordinated workflows and connected trading partners to turn intelligent decisions into reliable action.
Across healthcare, leaders are under pressure to make data more usable, artificial intelligence more actionable and operations more resilient. Autonomy is increasingly part of that conversation: systems that can sense changing conditions, support a decision and coordinate action with less manual intervention.
The ambition is not unique to healthcare. A 2026 IDC study sponsored by supply-chain software company Kinaxis surveyed more than 2,000 supply-chain leaders across nine global markets. Only 6% described their supply chains as autonomous at scale, while 41% expected autonomous operations to become their core operating model within one to two years. The same research found that only 12% had AI-planning governance fully embedded, 52% cited trust in AI-driven decisions as a top barrier to adoption and 62% said better data quality and integration would accelerate AI investment.
Those findings should be read as a measure of ambition and readiness, not as a forecast that most supply chains will become autonomous on that timetable. They do, however, expose the gap between what organizations want AI to do and the operational foundation required to let technology make or coordinate consequential decisions.
Healthcare supply chain makes that gap especially visible because decisions span providers, suppliers, distributors, group purchasing organizations, enterprise resource planning systems, electronic health records and other trading partners. The limiting factor is not simply the sophistication of the AI. It is whether the surrounding network can provide enough trusted, current context for a sound decision and then carry that decision across the organizations responsible for acting on it.
Data exchange alone does not create autonomy
Healthcare has made substantial progress digitizing individual transactions and connecting systems, but many workflows between those systems still depend on people to bridge the gaps. A supply-chain transaction rarely ends with the exchange of one piece of data; information changes as a product is sourced, ordered, shipped, received, used, invoiced and paid for.
When those handoffs are disconnected, teams reconcile information manually, investigate exceptions and restore context that technology did not carry from one step to the next. At GHX, we think of the accumulation of that rework, exception chasing and manual reconciliation as workflow debt. The term is useful because the burden compounds: each new capability has to operate on top of older handoffs that were never fully connected.
For health IT leaders, that debt shows up as more than supply-chain inconvenience. Manual bridges may sit between ERP, EHR, procurement, inventory, finance and supplier systems, which means modernization projects inherit the same underlying inconsistencies in item data, location identity, contract terms and transaction status. An AI model can analyze the information it receives, but it cannot reliably reconstruct business context that was never standardized or exchanged.
Independent healthcare standards work reinforces that point. AHRMM defines healthcare supply-chain data standards as documented agreements governing the representation, format, definition, transmission and use of information among trading partners. Those standards include Global Trade Item Numbers for products, Global Location Numbers for locations and the Global Data Synchronization Network for standardized master data.
That common language matters because connectivity and coordination are not the same thing. Connectivity lets information move; coordination preserves enough context around the information for the next participant to understand what happened, why it matters and what action should follow. The more responsibility technology takes on in interpreting signals or recommending actions, the more important that shared context becomes.
AI raises the stakes for trusted, connected data
Trusted, connected data can give healthcare organizations visibility into what is happening across a workflow: what changed, where it changed and which orders, locations or patients may be affected. AI can then help interpret those signals, identify patterns, surface likely dependencies and support decisions at a speed or scale that would be difficult to achieve manually.
The quality of the underlying context matters more in healthcare because supply-chain decisions can affect product availability, financial performance and clinical teams relying on those products. If item identities do not match across systems, location data are inconsistent or a substitution decision is detached from clinical use, an apparently intelligent recommendation may still be operationally wrong.
The industry already has examples of the discipline required to avoid that problem. AHRMM describes the GLN as a globally unique identifier for physical, functional or legal locations, while GTINs uniquely identify trade items and GDSN provides a mechanism for synchronizing standardized product information among trading partners. These structures were designed for transaction accuracy, traceability and patient safety, but they also illustrate the kind of dependable master data AI needs.
As AI takes on a larger role in operational decision-making, organizations therefore need confidence not only in the model but in the data lineage behind the recommendation. They also need guardrails around what AI can decide automatically, which decisions require human judgment, how exceptions are escalated and who remains accountable when a recommendation changes an order, substitute or inventory plan.
The sponsored IDC research points to the same governance gap: 52% cited trust in AI-driven decisions as a top barrier to faster adoption, and only 12% reported fully embedded AI-planning governance. In other words, the movement toward autonomy is not only a model-performance problem; it is also a trust, data and accountability problem.
Coordination has to precede orchestration
A decision creates value only when the surrounding network can carry it out. That is where the distinction between coordination and orchestration becomes useful. Connected data show what is happening, coordination establishes enough shared context to understand what it means across organizations, and orchestration links that understanding to the workflows and trading partners responsible for acting.
Consider a product disruption. A provider needs to know which orders, departments and procedures may be affected; a supplier or distributor needs to communicate what inventory or alternatives are available; clinical and supply-chain teams may need to determine whether a substitute is appropriate; and ERP, inventory and purchasing systems must reflect the decision. If each party can see only its own portion of the event, people become the integration layer.
A more orchestrated environment can move some of that work from manual coordination into shared workflow. Signals can be connected to the orders, locations and products they affect, potential alternatives can be surfaced with the relevant operational context, and decisions can be routed to the people or systems responsible for acting. The objective is not to remove human judgment from clinically or financially consequential choices; it is to reduce the avoidable work required to assemble the context those choices need.
This is also where the idea of an autonomous supply chain needs discipline. Autonomy should not mean that every decision is delegated to software. A more realistic progression is selective autonomy: routine, well-governed decisions can become increasingly automated while high-impact, ambiguous or clinically sensitive decisions retain human review.
The existing enterprise systems still matter
ERP, EHR, inventory, procurement and financial systems remain central to healthcare operations, and an autonomy strategy should not assume they will be replaced by one new platform. The opportunity is to connect workflows across those environments, improve the quality of the master data they share and reduce the manual coordination that sits between them.
That is another reason standardized product and location identities matter. A transaction may be technically successful inside one system while failing operationally across the network because the supplier, provider and clinical environment do not identify the same product or location in the same way. Autonomy built on inconsistent identities would simply automate reconciliation errors faster.
The same principle applies to modernization. Moving an ERP to the cloud can improve infrastructure, but it does not automatically resolve inconsistent item data or disconnected supplier workflows. AI layered on top of a fragmented operating model may make individual tasks faster while leaving the cross-organizational process dependent on manual intervention.
The pathway to autonomy runs through the network
For healthcare leaders, the practical pathway to autonomy begins with three questions. Can the organization recognize an important change with reliable data, can it make or support a decision with enough context to trust the result, and can the surrounding network carry that decision into action without requiring people to rebuild the workflow by hand? If any one of those steps is missing, adding more AI does not create an autonomous supply chain.
No provider, supplier, distributor or technology platform can create that environment alone because the decisions themselves cross organizational boundaries. The industry’s progress will depend on whether trading partners can align product and location data, preserve transaction context, create shared exception workflows and establish clear accountability for increasingly automated decisions.
Healthcare therefore should not measure progress toward autonomy by counting how many decisions AI is allowed to make. A stronger measure is how much reliable, repetitive coordination the network can perform without manual reconciliation while maintaining the human oversight appropriate to the decision. That is the difference between automating isolated tasks and changing the operating model.
Autonomy may be the destination, but trusted connected data provide the signal, coordination provides the context and orchestration turns a decision into action across the network. Healthcare will be ready for a more autonomous supply chain only when those foundations work together reliably enough that technology can act without forcing people to repair the handoffs underneath it.
Chris Luoma is Chief Strategy Officer of GHX, a healthcare supply-chain technology company whose network connects healthcare providers, suppliers, distributors and other trading partners.