Salesforce’s Pitch On Healthcare AI

ORLANDO, Fla. — Google and Microsoft made buzzy announcements this week at the HIMSS conference in Orlando, Florida around new generative artificial intelligence products and efforts to stand up guardrails around responsible AI use.

Cloud giants are currently jockeying for partnerships with healthcare organizations, especially hospitals, to aggregate their data and build AI programs from them, turning a hefty profit on the way.

Also active in the fray is Salesforce, a smaller cloud provider — but one with a major focus on healthcare, Sean Kennedy, Salesforce’s general manager of global health strategy and solutions, told Healthcare Dive at HIMSS.

In the lead-up to the conference, Salesforce announced a number of new AI tools, including conversational AI assistant Einstein Copilot Health Actions that clinicians can use to send referrals, book appointments or edit care plans. Salesforce plans to have the assistant in market this winter.

Salesforce is also currently piloting a program called Assessment Generation, which standardizes health assessments and uploads that information into the cloud. Assessment Generation will be available this summer.

The tools link with its Health Cloud, which builds on Salesforce’s massive customer relationship management platform — a key differentiator for Salesforce compared to Google or Microsoft, Kennedy said. The executive shared what sets Salesforce apart from its cloud competitors, along with future AI priorities and how the government should regulate responsible AI.

This interview has been lightly edited for brevity and clarity.

HEALTHCARE DIVE: What differentiates Copilot from other clinical assistants in the market?

SEAN KENNEDY: It starts with what differentiates Salesforce generally, right? We’re not going to sit in the clinical encounter between the doctor and the patient. That’s the electronic health record, that’s their space. A recording of that encounter is important. That’s not Salesforce.

But everything it takes to get the patient to that encounter — financially cleared for care, medically ready to engage in care — meaning they’ve been able to have a good experience, get their questions answered, schedule an appointment, be reminded to come in, complete eligibility checks and prior authorizations, we can do that whole journey.

So now, the clinician isn’t burdened with all this administration that they’ve had to deal with previously. Then after the encounter, it allows them to coordinate care on a longitudinal, ongoing basis. That’s where Health Cloud comes in. And that’s where Copilot is going to come in.

What about future use cases for AI?

We’re starting around the patients. We’re going to help take actions for patients and members as the consumers of healthcare, and make our care coordinators, call center agents, much more productive.

Then we’re going to move over to our providers in a September-October release. We’re going to work on, how do we establish better relationships with providers that are in a community outside of the organization, or among those inside the organization? So you can strengthen those referral lines. You can have a much more knit-together clinically integrated network.

So that’ll be the next one from a generative perspective. And then there’s a range of others. But a lot in the conversational AI space.

Within that, remember, you’re doing summarization and generation of information. You’ve got this concept of what we call “actionable analytics,” where you can ask questions to receive data and immediately take action on it. So, “Tell me all males over 50 that have not yet had a colonoscopy.” It’ll give you that list. And then you can run a campaign around that and get scheduled appointments.

So now you’ve got turnkey, clinically informed marketing campaigns with attribution that all started with that generation of a cohort. That’s the unique thing. No one does that. That is impossible for any other vendor to do today.

That sounds similar to Google’s search-and-summarization capability. What differentiates Salesforce here?

There’s a number of people that are providing generative capabilities. They may be able to summarize a segment like we can. The difference is the way we’re sourcing and unifying data, and then creating that segmentation.

Salesforce, at its core, is a multi-tenant platform on a single codebase. This is very different than a lot of other organizations that have multiple tenants on different platforms. We have updated our entire back-end to maximize our ability to take more action downstream. That’s different than what Microsoft is doing. It’s different than Google.

What’s unique about us is our ability to take action on it using our customer relationship management capabilities, through our marketing cloud, health cloud and others. And those aren’t things that they have, because they don’t have that capability.

Would that be your elevator pitch to a healthcare executive interested in partnering with a cloud company for them to go with Salesforce as opposed to Google, Azure or AWS?

What’s the outcome you’re looking for? If you’re looking to increase patients’ arrival rate, then you want to have a list of patients that may be at risk of not arriving. And you want to be able to send a message to notify them in a way that they expect to be notified. Some want a phone call, some want a text message, somebody an email. Sometimes you have to do it multiple times.

When they come back in and schedule that appointment, and they arrive, it goes back to that attribution — where that initial campaign that came out of the initial generated segments comes all the way back around. So now you can get an ROI on your efforts to action against that campaign from a population perspective. That round trip — that’s darn near impossible for anyone else to do without multiple vendors.

How much of a priority is healthcare for Salesforce?

Huge priority. It’s the one of the top two industries at Salesforce — financial services and health are the biggest industries we have. And we operate globally, and AI is in everything that we do.

That’s why I say you can’t look at Salesforce as just the generative AI widget. It’s the platform that has been purpose built to support where we’re going with AI, sourcing data, unifying data, actioning the data. And then generative AI is going to provide us that capability across those segments. So it’s squarely in our strategy for how we’re going to win in healthcare.

As a developer, what would you want to see from the federal government as they consider more regulations around the technology to ensure they’re not stifling innovation?

It’s a balance for sure. We were at the White House when Biden signed the executive order. I think that it’s important that we establish these guardrails around the use of AI, it’s really important that we have that human in the loop.

If I had a magic wand, what I’d want from the government would be transparency and openness. I think platforms need to be much more open, not only open with their application programming interfaces and their algorithms, but their documentation and organization. Salesforce — all our documentation is published online for folks to see. Our algorithms are going to be open and available for folks as they emerge.

I don’t think we can say that about a lot of other organizations in healthcare, but that’s the way we’re going to operate.