Artificial Intelligence At Cardinal Health: Two Use Cases
💡
Key Takeaway
Cardinal Health cut unplanned hospital admissions by 12.3% in oncology and scaled its AI-powered customer assistant rapidly enough to replace a shuttered call center during COVID-19 – two results that came from embedding machine learning into existing clinical and operational workflows rather than standing up standalone AI experiments. The patient outcomes improvement worked because the system pulled SDoH data from sources physicians rarely consult, giving care teams risk signals they could act on before a crisis hit; the virtual assistant worked because Cardinal Health had already integrated IBM Watson into its IT infrastructure, so scaling to customer-facing queries was an engineering decision, not a reinvention. Senior leaders should note that both wins required connecting AI to data and processes already inside the organization – the return came from integration discipline, not from the AI tool itself.
Cardinal Health cut unplanned hospital admissions by 12.3% in oncology and scaled its AI-powered customer assistant rapidly enough to replace a shuttered call center during COVID-19 – two results that came from embedding machine learning into existing clinical and operational workflows rather than standing up standalone AI experiments. The patient outcomes improvement worked because the system pulled SDoH data from sources physicians rarely consult, giving care teams risk signals they could act on before a crisis hit; the virtual assistant worked because Cardinal Health had already integrated IBM Watson into its IT infrastructure, so scaling to customer-facing queries was an engineering decision, not a reinvention. Senior leaders should note that both wins required connecting AI to data and processes already inside the organization – the return came from integration discipline, not from the AI tool itself.
Originally reported by Emerj Artificial Intelligence Research. Read the full story here.