AWS Summit DC 2026 | Faisal Sublaban, Sharecare

Clips
More from AWS Summit DC 2026

Faisal Sublaban

EVP of AskMD

Sharecare

play_circle_outline AskMD on Amazon Bedrock: 50-Year Clinical Reasoning Engine Democratizing Health Intelligence via a Unified Patient-Centric Data Layer

play_circle_outline Agentic AI orchestration for healthcare navigation, not prescription telehealth

play_circle_outline Deploying Multimodal, Multi-Model Clinical AI on AWS Bedrock: Evaluation, Benchmarking, and Distillation for Cost-Effective Safe Inference

Info
Transcript

Faisal Sublaban, Sharecare

Faisal Sublaban

EVP of AskMD Sharecare

In this interview from AWS Summit Washington, D.C. 2026, Faisal Hanafi, director of worldwide public sector at Amazon Web Services, joins Taylor Johnston, president and chief executive officer of University of South Florida Institute for Applied Engineering, to talk with theCUBE + NYSE Wired's Gemma Allen about how academia and cloud infrastructure are converging to accelerate defense-ready research and mission-critical innovation. The discussion centers on a newly signed memorandum of understanding between AWS and USF — grounded in Tampa's unique position as... Read more

explore Keep Exploring

How does AskMD bring a decades-old clinical reasoning engine to Amazon Bedrock to provide personalized, context- and insurance-aware health guidance and resolve fragmented health data so users can take the right next actions? add

Is AskMD a telehealth service (i.e., does it provide prescriptions or clinical outcomes)? add

How do you approach selecting, hosting, orchestrating, and evaluating AI models for healthcare applications to balance cost, safety, and operational flexibility? add

bolt Powered by CUBE AI

Faisal Sublaban, Sharecare

search

Gemma Allen

>> Welcome Matt to theCUBE here on the ground in Washington, D.C. for AWS Summit 2026 where we're having conversations across all things AI, agentic and the world of government and public sector. Joining me now for a conversation on healthcare, a very interesting space, one that we all hope can stay at pace with innovation and trust at the same rate is Faisal Sublaban, EVP of AskMD at Sharecare. Welcome, Faisal.

Faisal Sublaban

>> Thank you so much for having me. What an exciting time and, man, it's a bit chaotic over here. Lots of people.

Gemma Allen

>> I know. It's wild, right? There's a lot of folks here today, a lot of great conversations. So healthcare.

Faisal Sublaban

>> Yes.

Gemma Allen

>> We all assume that it's an area that's ripe for opportunity, accessibility, democratization, all that good stuff that AI has promised.

Faisal Sublaban

>> Yes.

Gemma Allen

>> It feels as though healthcare is the one space that everyone can get behind in way of what it promises.

Faisal Sublaban

>> Exactly.

Gemma Allen

>> You guys have done something interesting in that this is essentially a 50-year clinical reasoning engine that you're now running on Amazon Bedrock.

Faisal Sublaban

>> Right.

Gemma Allen

>> So that's taking something that's been long term trusted and bring it into the agentic AI era. Talk to me about that. Break that down.

Faisal Sublaban

>> Well, I mean, when you start to look at it, we're kind of shocked that something like this doesn't exist. When you start to look at when people go and do health related searches like Google, there's over 30 billion health related searches a month. And let's say we both are going there and we're asking a question saying, "My heart's racing." Well, it's kind of reckless to receive the same response back to both of us without having insurance-aware, understanding what your benefits are and coverages, not having context-aware, permission context around my medical records and medical history and labs and then understanding what's the action for me to ... What's the best next step and the action for me to take. And so AskMD does all three of those things. I don't believe that we have a health data issue. We have a fragmentation component. You have insurance information, pharmacy costs, provider information, your labs, medical records, it all sits in different portals. And so what we're doing is we're enabling a health intelligence layer in order to provide users access to all this information and let them have agency over their health.

Gemma Allen

>> But it's not telehealth per se, right?

Faisal Sublaban

>> No.

Gemma Allen

>> It's not giving you a prescription or an outcome.

Faisal Sublaban

>> No.

Gemma Allen

>> It's the navigation side of your own health.

Faisal Sublaban

>> Exactly. Yeah. And what we think is there's one data fragmentation issue and then there's a navigation component that are critical to solve. And people come in and healthcare is overwhelming. And when you don't understand what to do ... I always go back around AI and as you were speaking about healthcare being ripe for disruption and democratization and everything else, I go back to, with AI we can have it do what? You could write code, you can create and deploy an app or a platform the same day. You can have it create a 10-day itinerary to your favorite dream vacation to Tokyo, including making restaurant reservations for you. You can even have it write a poem for you in Snoop Dogg's voice. But when your kid wakes up at 2:00 AM with a fever, you're stuck with 1-800 numbers, portals that you don't remember the passwords to, or an ER wait time that could be two, three, four hours, and so that's the gap that AskMD fills.

Gemma Allen

>> So talk to me about the before and after experience, because folks are very familiar with AskMD. It's been around for a while. People understand the premise. What does this agentic option or this agentic layer look like? It's obviously more navigational device, but what exactly does it offer consumers and users and enthusiasts? And so there's many-

Faisal Sublaban

>> At Sharecare, we've had over 15 years with over a billion dollars spent on capabilities. And the same thing that ends up happening is we end up having data that sits in isolated areas and not in a language that is useful to a user. So just because I could connect my medical records into the platform or my labs doesn't mean I even understand what it means. And now with AI, that's what it's able to do. And so from our standpoint, our orchestration layer and our governed orchestration layer, that's the product. The model is not the product. And ultimately, at the end of the day, it's how do we use agentic workflows around healthcare in order to take specific modules and data points in order to more seamlessly provide a translation and a unified language that becomes useful to the end consumer.

Gemma Allen

>> So talk about AWS. We're here at the AWS Summit in D.C. You guys are obviously a key partner. Talk about that relationship and the evolution of Bedrock as the kind of underlying architecture for this.

Faisal Sublaban

>> Yeah. So we take a multimodal model approach. All those models are retrieved and accessed within Bedrock. The Bedrock we also utilize. This is healthcare. And so, one thing when you start to look at guardrails, security, privacy, these are non-negotiables in healthcare, and being able to leverage the Bedrock platform in order to do so and have guidelines that we're able to continuously have discipline around deploying is critical. Also, when you look at our agentic layer around our orchestration, the agentic workflows all run through Bedrock and it's a critical component for us to ... When you take a modular approach to the architecture in order to activate those modules.

Gemma Allen

>> So people say all the time Bedrock's main advantage is kind of LLM Frontier Lab agnostic, right?

Faisal Sublaban

>> Yeah.

Gemma Allen

>> You're not getting into bed with anyone, you have the Switzerland approach, but we know that preferences are built and that solutions are better solved with particular outcomes and inputs in case of the Frontier Lab conversation. What are you seeing across these models? Are there models you're seeing used more in healthcare? We hear a lot about Anthropic and enterprise is there. .

Faisal Sublaban

>> Yeah. I think it's the same cast of characters that you see, Anthropic kind of spearheading the way. But from our standpoint, taking a multimodal approach, we also are able to host and have on-prem our own clinical reasoning model. And so for us, when we start to look at the orchestration and how we direct that traffic, it may be cost prohibitive to use one of these larger foundational models and we may need to use a distillation model. We also were able to bring our model, host it within the AWS infrastructure, run it through Bedrock as well. And so it's that flexibility that you need to have, especially around healthcare and how fast this technology is moving. And from our end, it truly is how do we provide guardrails and an evaluation framework that enables us to have confidence and a benchmarking tool and mechanism that allows us to bring in and introduce new models after they've been tested within our workflows.

Gemma Allen

>> So it gives the level of stability, predictability, and security, I guess, which is key, right?

Faisal Sublaban

>> Correct.

Gemma Allen

>> Parameters.

Faisal Sublaban

>> Yeah.

Gemma Allen

>> When we think about the competitive element to this, we all had a world where people were experts because they used Google, right? I had a GP one time say to me, "Please don't Google this."

Faisal Sublaban

>> Yeah, I get that question all the time.

Gemma Allen

>> We know how hypochondriac this is too for some, right?

Faisal Sublaban

>> Right.

Gemma Allen

>> But talk to me about competing with just folks using LLMs to get self diagnosis and must be happening at a scale that's unheard, right? When we think about how easy ... You mentioned parents, for example, parents having a fever. The temptation now to ask ChatGPT, "What's wrong with my kid?" It must be stronger than ever before. How do you think about that from a competitor for a second?

Faisal Sublaban

>> Yeah. So I think that there's a couple things. When you look at context-aware, that's critical. And then how do we provide agency to a user around their health? And this is not just when it's episodic and you're going through an emergency or your kid has a fever, but more day-to-day engagement. And so what we look at is the commitment curve. You look at the first time that you went into ChatGPT and you asked it just a simple question, answer that came back and then your mind starts racing of all the different things you've always wanted to ask that you couldn't do the research on, some maybe appropriate or not, or things you feel silly for asking. But at the end of the day, it's once you got a taste of intelligence, you started to dive deeper. And so our premise is that once we expose users to health intelligence and the ability to have our orchestration layer bear all the weight and complexity of healthcare, we find that it's going to provide a ton of utility that's truly differentiated. And at the end of the day, starting that democratization and empowering the user to not just understand their health, understand what they should be doing, how they engage in their health, both daily and when there's issues like emergencies that I reference. And then at the next component of it is you reference providers and doctors where the doctor you said was telling you don't. I mean, we look at it totally different. We look at it as imagine if you were a doctor and you had a patient that was more educated, understood what was happening within their health and could clearly articulate that to the provider. Imagine how more impactful they would be instead of, "Tell me what's going on." And then starting from fresh, when you look at the search manage assess approach that we take, that's the utility we provide both to the user and also to providers.

Gemma Allen

>> I mean, I think there's a lot of old cranky Irish doctors that probably are not hoping that happens, but I think it's a great idea from the perspective of society, right?

Faisal Sublaban

>> Exactly.

Gemma Allen

>> So at the core of this though, it's trust, right? It's that these are trusted. This is a trusted space, you're not going to be fed false information or melodramatic information or any of that stuff, which can sometimes, again, drive traffic in other directions, right?

Faisal Sublaban

>> Right.

Gemma Allen

>> How do you advocate and kind of use that from a go-to-market perspective in a landscape where everything's a bit noisy and there's almost too much information available?

Faisal Sublaban

>> Yeah. Well, I mean, I think it first starts with identity management and that's where once you start to have medical records and other health information connected such as your insurance or benefits, labs, records, provider information, we necessitate that we do identity proofing with IAL2 certification. And so we've partnered with CLEAR and ID.me in order to provide that extra layer of protection in order to ensure you are who you are. That's the first layer of trust. And if you're not going through biometrics in order to do identity proofing, we believe to be that's pretty reckless, especially with your most proprietary information, your health. You don't want your medical records floating around everywhere about every ailment or issue you may have had or not had, and so that's the first part of what we do. And then from there, it's all permissioned context. So the user controls their data no matter what. So if you want to ask about a specific condition around your torn ACL, it doesn't mean that you have to enable access into your complete medical history over the last 30 years. And so that's the other component and we feel as if we evangelize agency over your health, the user is at the center of it and they control their data. It's a pretty powerful message.

Gemma Allen

>> It certainly is. Talk about the relationship with AWS from here. I mean, there's a lot happening here, a lot of announcements. What those announcements last year or two is most relevant to the healthcare space, and what's ahead? What does this relationship look like a year from now?

Faisal Sublaban

>> Yeah. I think the business of working with AWS, our relationship started like many just as far as infrastructure goes and then now it's starting to expand. And as they bring more products to the table, there's also opportunity costs. We want to focus around the critical components and the proprietary orchestration layer that we have and being able to build that out and provide a ton of utility to our users. We're not looking at a lot of the capabilities that AWS is able to natively unlock within the Bedrock solution. We want to leverage their scale. Additionally, you get to use the power of what Amazon brings to the table. They have a lot of super brains, as I call them, that maybe companies don't have on their bench that you have access to and AWS really takes a consultative approach to try and help you find the next best action, but enabling you to take your platform, scale it to where it's safe, trustworthy, and there's actual sustainability around it.

Gemma Allen

>> So it allows, I guess, folks to meet the moment where it's at, right?

Faisal Sublaban

>> Exactly.

Gemma Allen

>> Which is so important, especially in healthcare.

Faisal Sublaban

>> Exactly.

Gemma Allen

>> Faisal, thank you so much for joining us on the theCUBE.

Faisal Sublaban

>> Oh, thank you. Yeah, appreciate it. Thank you so much.

Gemma Allen

>> I'm Gemma Allen here with theCUBE at the AWS Summit in Washington, D.C. 2026. We are bringing you all the live action on the ground, some great conversations coming up, one of the head, global head of public sector for AWS. Don't go anywhere.