Topic
26 facts from 2 episodes, in date order. These are what speakers said on the shows, each linked to the moment it was said. They aren't independently verified.
Undated
Josh Weiner says CVS Health is one of the largest integrated health care companies in the US, with subsidiaries including the insurer Aetna and the pharmacy benefits manager Caremark.
“CVS is one of the largest integrated healthcare companies in the US, with subsidiaries that include Aetna, a health insurer, Caremark, pharmacy benefits manager, the largest retail pharmacy”
Undated
Josh Weiner says CVS Health is increasingly expanding into care delivery businesses such as Oak Street Health and Signify.
“as well as increasingly some care delivery businesses. You may have seen in the news Oak Street Health, Signify.”
Undated
Josh Weiner says that before AI, patients were assigned care plans based on their condition and sometimes gender, age and behavioral preferences.
“for a given person, based on a condition, and if you're lucky, your gender, your age and some understanding of behavioral preferences, you're assigned a care plan.”
Undated
Josh Weiner says AI lets CVS Health use long-term unstructured health data, including claims, electronic medical records, labs and wearable data, to personalize care.
“we can now make the most of our longitudinal set of unstructured health data, medical claims, electronic medical records, labs, as well as wearable data.”
Undated
Josh Weiner says CVS care management is typically provided by pharmacists or health plan nurses, who do not prescribe medication or perform surgeries.
“it's something that either a pharmacist from the pharmacy business or a nurse from one of the health plans typically provides. And they are not prescribing medication or performing surgeries.”
Undated
Josh Weiner says AI agents can share data across different systems without a multi-year, costly integration of legacy systems.
“we can work around different systems and different data models and have agents talk to each other to facilitate data sharing without spending three years and a lot of money trying to integrate legacy systems”
Undated
Josh Weiner says CVS Health's subsidiaries use different clinical platforms and data structures, which made building an internal customer 360 view difficult.
“We have all these subsidiaries with different clinical platforms, different data structures. So for us, developing even an internal customer 360 has had the issues you've just described.”
Undated
Josh Weiner says CVS Health holds a NoSQL vector database of electronic medical records that was previously impossible to interpret and share.
“We have a whole NoSQL vector database of electronic medical records. It was previously impossible to interpret, impossible to track the data, structure it and share it.”
Undated
Josh Weiner says large language models and prompt engineering now let CVS Health convert and share its electronic medical record data.
“Now you can just throw LLMs on top of it and it works. And we can use prompt engineering to figure out how to convert it and share it with all different types of recipients.”
Undated
Josh Weiner says CVS Health is putting governance in place to test, co-design and measure the accuracy and effectiveness of its AI experiences.
“We're putting in quite a bit of process and governance to make sure that we're extremely rigorous in how we test, how we co-design these experiences, how we measure accuracy and effectiveness.”
Undated
Josh Weiner says CVS Health's AI work starts with back-office and enablement uses and will expand to broader uses as success is shown.
“The starting point is back office and enablement, and it's only over time as we demonstrate success.”
Undated
Josh Weiner expects people to use a few agents from major consumer platforms and operating systems, with CVS agents integrating into them.
“I expect that Sam, you'll have a couple agents, predominantly, maybe from some of the major consumer platforms and operating systems you work with today.”
Undated
CVS Health hosted 25 small language models in its data center alongside large language models, which Yogesh Bhagat says cut the cost of its IVR system by five to six times.
“We hosted twenty five S L M s in our data center, working in conjunction with the L L M s. What it helped us, it reduced our cost by five to six x.”
Undated
Yogesh Bhagat says CVS Health's interactive voice response system handles more than a million calls every day.
“when we looked at our I V R calls, which is million plus call every day.”
Undated
Bhagat says the small language models also let CVS Health meet the sub-second response time that an IVR system requires.
“In addition, it also helped us to meet the sub second response time, which any I V R system would require.”
Undated
CVS Health built an enterprise AI development platform that provides security, guardrails and observability by default for every agent built on it.
“we actually build an enterprise A I development platform, which brings security, guardrails and observability by default for creating any agent.”
Undated
Bhagat says CVS Health validates its shared development platform once rather than reviewing each AI agent separately.
“So you validate the platform, not every agent which goes.”
Undated
CVS Health's approach, as described by Bhagat, is that no AI agent may launch as autonomous and each must earn autonomy by proving consistent behavior.
“Nobody will have a right to launch an autonomous agent. You have to launch an agent and earn the respect.”
Undated
Once a task such as a date-of-birth check is shown to work consistently, Bhagat says CVS Health removes the judge model and the human reviewer from that task.
“We remove the larger model out of the context, judge model out of a context, we remove human in the loop, and we know that it is always working fine.”
Undated
CVS Health's developer platform automatically creates a testing agent for each AI agent that is built.
“What we have done is every agent we build, our developer platform also builds you an agent to test it.”
Undated
CVS Health's testing agents are prompted to act as older patients with health issues to generate test questions and show frustration.
“We say, hey, agent, you are an O A P S person, old age person, and you have X number of basically healthcare issues.”
Undated
In CVS Health's setup, smaller models give the main answers and larger models cross-check them as judge models.
“your main models, which are generally a smaller models, are giving an answer. And then you are cross validating the output just by a larger models.”
Undated
At CVS Health, a repeated question or caller frustration triggers escalation from the small model to a larger model, and then to a human if needed.
“if the question is getting repeated twice, then I need to escalate to a larger model and validate my response.”
Undated
Bhagat says the human in the loop at CVS Health is the contact center agent who takes escalated calls, and those calls become feedback for training.
“Human in loop is the agents who are finally picking up the customer calls.”
Undated
CVS Health replaced a legacy tree-structured IVR menu with a conversational AI system.
“We were replacing the old legacy I V R, which used to say press one. It's a tree structure.”
Undated
Bhagat says running frontier models for the IVR would have kept the project at the pilot stage.
“Frontier models, if I would have run frontier model, then I would have not been able to make it live. It would have only be a pilot project, would have never gone live.”
Episodes