AI ‘smart agents’ underutilized in fight against healthcare fraud
Stopping fraud before it’s perpetrated, cutting illicit payments and reducing hours spent by staff on such cases : These are three of the top four benefits healthcare administrators feel AI “smart agents” could deliver to their operations.
Stated another way, only one of the top four benefits healthcare admins would expect from an AI smart agent didn’t involve fraud, waste or abuse.
That’s according to a survey conducted by the online media channel PYMNTS in collaboration with Mastercard’s AI company Brighterion.
In an April 15 post summarizing the results, PYMNTS defines smart-agent technology as a form of AI that “can be assigned to multiple entities within a system and offer highly personalized decisioning capabilities.”
In its new survey, the company further found that more than a third of healthcare managers are at least “very interested” in implementing smart agent-based AI, while 13% are “extremely interested.”
The authors of the report note that, in various other industries, machine learning and artificial intelligence “have yielded powerful tools to manage such complex matters” as fighting fraud, waste and abuse. “Yet these advanced computational systems have a long way to go in healthcare administration.”
Only 4.3% of the healthcare sector currently use AI, according to PYMNTS’ latest research. By comparison, close to 10% of financial institutions have adopted the technology.
AI applications stand to save billions in healthcare administration annually, the authors assert.
“Better fraud detection also reduces false positives, erroneous judgments that block legitimate transactions,” they add. “A system that effectively stops fraudulent transactions is better at letting good ones through, and institutions that adopt AI-based fraud detection platforms reduce instances of false positives by more than tenfold.”
Click here to read an earlier and more detailed summary of the same survey report. Click here to download the full report (contact info required).