What Is Agentic AI?
Like other forms of artificial intelligence, agentic AI is only as accurate as the data that fuels it. It relies on “a digital ecosystem of large language models (LLMs), machine learning (ML), and natural language processing (NLP) to perform autonomous tasks on behalf of the user or another system,” according to IBM.
Agentic AI is the term used to describe the overall concept. AI agents are the individual components within the model that are created to handle specific tasks and processes. Agents within an agentic AI system have the “agency” to analyze data and then make decisions about what to do with the results.
While a significant step forward, both Saunders and Jason Warrelmann, vice president of healthcare strategy at UiPath, caution that agentic AI is still considered artificial narrow intelligence. Artificial general intelligence, which would allow machines to think like humans, does not yet exist.
“Right now, the best we can do is provide context so that the agent understands how to answer. There’s still a large language model behind it, so the agentic AI isn’t acting completely on its own,” Warrelmann says. “The computing required for that is still beyond us.”
“While agents and reasoning are powerful capabilities, they’re still no match for the incredible complexity of human intelligence,” Saunders agrees.
EXPLORE: How can data governance and LLMs help healthcare organizations avoid bias and inaccuracy?
How Does Agentic AI Differ from Generative AI?
Generative AI applications use data from large language models to craft responses. The quality of the output relies largely on the specificity and guidance provided by the user, a process known as prompt engineering.
Agentic AI is more proactive. It can pull information from multiple sources, use sophisticated reasoning and then automatically complete the next task.
“Agentic AI builds on generative AI, taking simple responses further with the ability to consider options, go back and redo steps,” says Saunders. “It works much more like we do when we solve problems and work out how to consider new information.”
In healthcare, agentic and generative AI can work together to increase efficiencies and boost productivity. For example, after a surgery, generative AI can use the patient’s record and the surgeon’s notes to write post-op instructions for medication use, activity limitations and follow-up care.
Agentic AI can then share the generated instructions, monitor if the patient has accessed the document within the patient portal and send reminders about future appointments. If the patient reports a serious symptom, the healthcare AI agent could automatically alert a nurse or schedule a virtual consult with the provider.
