Why St. Luke’s Used Agentic AI for Supply Chain Disruptions

St. Luke’s Health System, a health system based in Boise, Idaho with 8 hospitals and nearly 400 clinics, was facing a significant challenge in proactively identifying and responding to disruptions within its supply chain. Disruptions can come from major weather events, item discontinuations, product recalls, low supplier stock, or a shortage of raw materials. These disruptions can cause delays in care if the right product substitute isn’t identified quickly.

The health system didn’t have a uniform way of managing disruptions. It relied on a combination of digital, analog, and informal methods that notified teams when there was a disruption, a time-consuming process that was hard to operationalize. St Luke's relied on email, spreadsheets, and manual back-and-forth to manage disruptions — an imperfect system that led to disjointed team coordination, repetitive work, and last-minute scrambles.

At the same time St. Luke’s Health System was dealing with a growing threat of disruptions, it was getting ready to open a 330,000+ square foot consolidated service center (CSC), putting supply chain distribution, pharmacy, and logistics under one roof. While this $71M flagship investment promised to offer St. Luke’s unprecedented control over its supply operations, it also meant the health system needed to manage the 7,200+ unique supplies across the massive logistical hub. 

With a new state-of-the-art CSC, the St. Luke’s Health team built new processes to manage and respond to supply shortages, backorders, and other disruptions. But with 57,000+ items, 20+ associates managing disruptions, and a complex, multi-step process that relied on close team coordination — disruption tasks would stall in the process. 

Disruptions can be a costly problem for health systems without the right tools to tackle them. Systems that move too slowly in the face of a disruption can be stuck with purchasing higher-cost, premium product alternatives, additional fees for rushed freight, and even risk of delays to patient care.

St. Luke’s Health System needed an AI tool that could identify when a disruption was occurring and help identify a clinically equivalent product substitute. It then was looking for a partner that could deploy agentic AI to help streamline the highly manual and labor intensive disruption response process. 

Alongside 10+ warehouse management technology partners, Clarium was selected as the partner of choice for control tower, supply disruption monitoring and response coordination. The task at hand: ensure St. Luke’s team is proactively responding to the supply disruptions that impact the system, and the CSC that now services it.  

Unlike other companies offering this service, St. Luke’s liked that Clarium started at the height of the COVID-19 pandemic with an AI-native platform. Other companies had just entered the space, cobbling new code on top of old code.

As a platform company, Clarium offered the potential to serve as a long-term partner with a concrete roadmap. 

First, Clarium unified St. Luke’s numerous disparate data sources from its warehousing, EDI, ERP, and other systems. This unified and cleansed data gave St. Luke’s greater transparency into supply chain operations than ever before.

From there, Clarium custom-configured its disruption monitoring and response coordination solution for St. Luke’s complex CSC environment. The two sides engaged in regular working sessions to iterate on Clarium’s AI agent, which automates task management so St. Luke’s teams move quickly and in lockstep coordination to resolve potential supply disruptions.

The Clarium Disruption Agent at St Luke's is a custom-configured task management and process orchestration quarterback. It intelligently guides the disruption event from task to task until a substitute item is identified and it's resolved — shepherding across sourcing, inventory, purchasing, and distribution teams so a disruption never slips between the cracks, teams stay on task, and the disruption is managed as quickly as possible.