Tailoring AI Strategy for Healthcare Leaders: Product Adoption and AI Literacy | Ritu Chakrawarty

AI strategy for healthcare cannot be copy-pasted from other industries — Ritu Chakrawarty on AI readiness, product adoption, and AI literacy for healthcare leaders.

AI strategy for healthcare cannot be copy-pasted from another industry. Ritu Chakrawarty, an analytics and AI product leader with nearly 2 decades across analytics, consulting, startups, and enterprise AI strategy, argues that the reason most healthcare AI projects stall is not model quality but adoption design. Recorded at the Put Data First Conference in Las Vegas, this conversation lays out a practical adoption-first framework for healthcare leaders who are tired of pilots that never scale.

What We Cover

• The "stop, easy, and flow" framework for designing AI that actually gets adopted: identify what people will stop doing, make the solution easy to use, and embed it in the workflow where people already operate.
• The "should or could" test that separates problems that need AI from problems a simpler solution would handle faster and cheaper.
• Why AI literacy has to be sponsored from the top of the organization, and why fear of AI is fundamentally a fear of not knowing what is coming and why.
• Why every CEO needs an AI advisor reporting directly to them, and why the right advisor blends technology depth with business acumen rather than being purely technical.
• A litmus test for AI viability: start from the value, verify the information is digital and verifiable, then test for adoption before scaling.

Key Takeaways

AI adoption is a product problem, not a model problem. Most healthcare AI initiatives fail at the adoption layer. The technology works; the humans around it are not ready to change their workflows.

"Should or could" is the most underused question in AI strategy. Leaders regularly greenlight AI for problems that could be solved with a form, a rule, or a spreadsheet. The real cost of AI is not the model; it is the organizational debt of maintaining it.

A purely technical AI advisor creates a bias toward technical solutions. The best AI strategists are generalists who have delivered business outcomes across multiple parts of the organization and bring strong technology understanding to the conversation.

Frameworks & Tools Mentioned

• Stop, Easy, Flow adoption framework
• Should or Could AI viability test
• AI literacy sponsorship model (top-down)
• CEO AI advisor role

About Ritu Chakrawarty

Ritu Chakrawarty is a data and analytics executive with nearly 2 decades of experience spanning analytics, consulting, startups, and enterprise AI product strategy. She advises leaders on how to move AI from pilot to adoption and frames AI literacy as an organizational capability rather than a training program.

Related Resources

Related episodes:

• From AI Strategy to Execution: Trust, Leadership, and Operational Reality
• Data Readiness for AI Adoption
• From AI Hype to Real Value

Related topic: Healthcare AI Strategy

Related article: Healthcare Data Strategy

Stay connected: The AI Health Pulse newsletter

Enjoying the show?

Honest ratings on Apple Podcasts are the single best way to help more healthcare A

📘 Beneath the Signal, Chris's book on the human work behind trusted data and responsible AI in healthcare: Get it on Amazon

Transcript

Copy Transcript Copied

Ritu Chakrawarty: 0:19

It is because people are not really adopting the product. It is not in the flow. It is not easy for them to use it. Sometimes it is just the status quo. For every company, the test is the same.

Christopher Hutchins: 0:37

Good, on location here at Planet Hollywood in Las Vegas for the Put Data First Conference. We are having a blast, and today I'm really excited to be joined by Ritu. Welcome to the Signal Room.

Ritu Chakrawarty: 0:48

Thank you.

Christopher Hutchins: 0:49

Thank you so much for agreeing to just sit down and have a conversation. I think it's been remarkable how much passion and enthusiasm we're running into in this room over the last couple of days. And getting to meet you has been wonderful and meeting so many different people. It just seems like we're at a really different point because when you're at the beginnings of a transformation of any kind, people tend to be a little tentative and kind of laid back, kind of waiting to see how this is going to play out. But everyone's leaning in and they're so excited. So just to start with, tell me a little bit about your background. I know you're doing AI strategy and pretty cool stuff, but I don't think I've seen you without a smile on your face since I saw you yesterday for the first time. So you clearly are passionate about what you're doing. Love to understand what's behind all that.

Ritu Chakrawarty: 1:38

Absolutely. First of all, thank you so much for having me here. I am glad that I could make it. I was kind of skeptical about taking two days off, but I think I made the right decision. There was so much energy in the room to learn from each other, how things are happening. Starting with my background, I'm in this industry almost two decades, so very long. My degree, I'm an engineer, and my first taste of AI, I would say, was my research, which I did in neural networks. And during that time it was very specific use cases, like how do you do digital signal processing? That was my PhD paper in engineering. And then that was also a time when software was taking shape. Started my career as an analyst, and from there, I would say through about 12 years, during that time I left the neural network space, but then came big data, and that was the talk of the town. Since then, I was part of execution strategy in consulting for applying intelligence, did my startup for a while, and then I realized that I should be in enterprise doing something more in the AI space. So now I'm working on AI product strategy, putting together the AI product for commercial use.

Christopher Hutchins: 3:23

That's exciting. How are you finding it? The challenge that I've been running into because I've been on the chief data officer side in the healthcare sector. There's so many solutions coming at us from every direction. Generally, they are addressing a really small part of a very large gap that we try to fill. And the difficulties with that are primarily financial ones because you can't afford to have 10 things to fill one gap. And then sometimes it's not even the right gap that we're trying to fill. So it's complicated, but in your job, what you have to think about is that not only is it the right solution, but strategically, when do you do certain things and when is it going to be meaningful? When is adoption going to be a problem? When is it a risk for a compliance issue? So many things that go into what you're responsible for. I'd just love to hear what your approach is.

Ritu Chakrawarty: 4:12

Right, absolutely. And as you said, there's so many solutions. Every day you face something else, and then you wake up and you find another new entrant in the market. So for a person like me who is responsible for putting together what the product strategy looks like, finding how should we build what is already in the market, or just latch on to something which is already there and build on that. So that is the build versus buy decision, which is very difficult.

Christopher Hutchins: 4:46

Right.

Ritu Chakrawarty: The way I approach it is starting from the value, what we are trying to achieve, where does the value lie? It looks complex, but looking from that angle, where is it going to increase our revenue, where is it going to give us a cost savings? Starting from that angle, then identifying which are the areas where you can have these values quickly achievable? The low-hanging fruit. And then I apply a litmus test for that. Very simple litmus test. Anything where AI could help us, before even finding the right solution, first of all, is it an AI solution? Does this information exist digitally? Can I verify it? Because if you can't trust it, you can't use it, you can't adopt it. So this is my first test for adoption. Once I have that quick list, a litmus test on the viability as well as the feasibility side without involving too many people, then it comes to bringing the right people in because what I always say is that a product fails not because the LLM or the model is bad. It is because people are not ready to adopt it for multiple reasons. It is not in the flow, it is not easy for them to use it. Sometimes it is just the status quo, they are very comfortable in their environment. So bringing them in from the beginning. I use a very quick framework: 4:47

stop, easy, and flow. Basically, if I bring this product, what are the three things you are going to stop doing? That helps me bring them into the mix and drive adoption from day one. And then during the design, you design the product in such a way it is in flow. So it's not another point solution. For example, if I'm analyzing some information and I'm in Excel, I don't want to go into another chat box to identify what's happening in the data. I want to know, hey, here is the 3% variance. Why is it happening? Can I do something about it? Can I have a conversation about it? And that's where it is easy to use. Similar thing if I'm working in SharePoint and I have a lot of content. I want to know in the last month, what are the different things happening? Can I have a summary and just send it to my boss? I need to have that conversation right there. So that's the flow. And then the easy part is, is it easy for me to navigate or do I need to do three or four steps, enable certain things, and only then can I use my product? So these are the three things on the product designing side which have helped me not only bring a product which is useful, but that people want to use.

Christopher Hutchins: 7:55

I think AI is getting the same kind of interest and reactions that dashboards got at one point in time. It was amazing when we first started deploying them, in my first experience the very first one just caught on. People got excited, and all of a sudden they had dashboard envy. You happen to have a very well thought out problem that you wanted to solve, and the dashboard seemed to be the right thing. We work with you, you launch it, one of your colleagues sees it, they're like, I want one. And they call me, tell me they want a dashboard, and my team goes through it, and it's like, they really just need a list. But then the worst part is the over-engineering. The people that design some of the solutions are incredibly brilliant people. What motivates them and makes them tick is creating something that's really useful and powerful, but they can get really excited and over-engineer it. And I'm not nearly as technical as most, but I have tended to over-engineer some things myself. But the problem is if it's not explainable, it's like a comedian telling a joke that no one understands the punchline. If it's not self-evident, you didn't do it right. If you're not answering the very first question that someone needs to have answered, is it going to enable a better decision to get made? Is what you're showing me going to change any decision that I can make at all?

Ritu Chakrawarty: 9:18

Improving my agency, right?

Christopher Hutchins: 9:20

Right.

Ritu Chakrawarty: 9:20

Helping me to do something better. And that's where it actually helps. Because the easiness, it's in flow. I'm not taking three extra steps. Because as long as you ask people to do additional things, they say, I'm already overloaded. I already have so much on my plate, and you are asking me to use this new tool. I have to do three steps. No, thank you so much. I am gonna use the same thing I've been doing. So, yeah.

Christopher Hutchins: 9:46

You know, it's interesting. The quick reaction sometimes, because everyone's excited, is they want to come up with a new solution. But one of the most glaring solutions that I saw in the last few years was at my doctor's office. And what they were doing that I'd not seen people doing well was so simple. I'm walking out the door to the waiting room, and she made the appointment for me. Right then and there, took her 10 seconds, and she handed me a business card with the schedule on it. Gaps in care is a real problem in healthcare because the basics of clinical protocol are not pushed down to the levels they should be. So the physician knows what he needs to see the patient, but he doesn't do the scheduling, doesn't probably even have access to the schedule. But it was really as simple as in that particular practice and specialty, the clinicians made sure that the front desk people who are interacting with the patients knew what the standards were and took care of it. A piece of paper is something that people are trying to solve with a whole bunch of technology now.

Ritu Chakrawarty: 10:43

Right.

Christopher Hutchins: 10:44

But the right solution is sometimes the really easy one.

Ritu Chakrawarty: 10:48

I always say that just because AI is a shiny tool, it's not necessarily an AI solution. So do the should or could test. Should I be using AI, and then it comes to could, can I do it? So that's where the litmus test is. First is the viability about should, and then could is, do I have digital information? Is it right? And also, if it is very easy for humans to do, like you gave as an example, I shouldn't be making it complex and putting it into a system until I see that I need to scale it or extend it. So yeah, this is part of AI product strategy.

Christopher Hutchins: 11:31

Well, the good news is that people that are really smart and can do those things, they have you. So you keep them grounded. Hey, slow down. I know you like to use your hammer, but we can actually use a piece of tape on that. It's really not that serious.

Ritu Chakrawarty: 11:45

Absolutely. And that's exactly what happens because sometimes you don't need a full big solution. You just need something simple. And that's where, as you rightly said, this is such an important discussion because many times, especially if you're working with a tech team, they tend to want to build big.

Christopher Hutchins: 12:03

They're great at it.

Ritu Chakrawarty: 12:03

Where they want to say, oh, let's have a 20-member team, then let's build it. And then my question is, wait a minute, within our existing tools, you may have certain capabilities which you can immediately start using and at least have a 30% benefit. Can we use that and then see the value and adoption? Are people using it? That is a signal. If people are coming back to your solution, even if it is halfway, then you know that they like it, and then you want to make it better and better. So first find out whether they want to go from point A to point B. Have that cycle and then build the Ferrari after you see that they are taking that journey.

Christopher Hutchins: 12:50

The interesting dynamics tend to be cultural in a lot of ways. We've talked about this in so many different contexts over the last couple of days, but human trust is kind of a big deal. There's fear involved, people are afraid it's coming for their job. When you're working with teams and identifying the workflows that you're going to go after, oftentimes the experience has been that, oh, here they come, they're gonna ask me to do something else that's gonna change my workflow. And to your point, it's not in flow. What are some things that you're dealing with trying to actually counter what those experiences have been in order to get them to be comfortable and willing to trust that she's not coming for my job, she's trying to help me become more valuable and work on higher priorities?

Ritu Chakrawarty: 13:38

This is such an important thing. And I think everyone in our position needs to think about this. So when I use the framework for stop, easy, and flow, the most important thing is then what you start. So you're stopping something, what are you starting? Rather than telling them what they're going to start, my question is, what is this going to enable you to do that you want to do today but can't? And as soon as you say that, that's where you are building their trust because now they know that okay, this tool is coming to help me be a better version of myself. Improving my capabilities. That's where you actually latch on and say, your job is not going anywhere, rather, you are opening yourself for better things. And I like the way I was reading about how most of the jobs that will change with new tools, including physical AI, are jobs that people don't even want to do. If it's a dirty, dangerous, or difficult job you're taking away from a human, that's helpful. Because I don't want to go into a ditch to do some kind of repair. That's difficult. I don't like to do it. But if something can do it and I will just operate from my desk and move the robot, that's what I would love to do. I'm not losing my job, but I'm getting a better version that helps me not put myself at risk.

Christopher Hutchins: You know, it's interesting working around data analysts, data architects, ETL developers for a number of years. One of the things that I've observed, I'm sure you've seen it too, analysts spend easily 60 to 70 percent of their time wrangling the data, getting it fit for use, very little time doing actual analysis. And when you actually can automate those things, they're a little nervous about it to begin with because they see their value in a different way. So I've always had to coach my team: 15:31

you're not here because you can do stuff with technology.

Christopher Hutchins: 16:04

You're here because you're able to figure out what's the right solution to the problem that we're trying to solve. And it's what's in your head, what you know, how you think, how you approach things, how you network with people. That's the stuff that really is valuable. And it can actually start to restore some trust with the people who, to your point, are fearful of losing their jobs.

Ritu Chakrawarty: Also, I would say the most important thing: 16:31

fear is darkness. Fear is when you don't know.

Ritu Chakrawarty: 16:40

When you don't know what is coming and why it is coming. So start with educating your people and that's why AI literacy is super, super critical. And it has to be sponsored from the top. As long as people are able to know what it is and why it is, they will find their own path. But if they don't know, they will speculate. Because even for good AI to work, as you said, you are here for a reason because we know that data works in a certain way. You have institutional knowledge, you have that recipe for how people work. That's not going away. And as long as people understand that, that's where they will start leaning in. That's where they will start adopting it. Then whatever barrier you see is slowly loosening up. And that's where you start with literacy and make it visible from the top. For example, you record a video with an AI tool and then share with your team saying, I'm super excited, I got this tool, I made this video, and this has helped me to do the thing which used to take five takes, now it's done in one take. Tell me what's your feedback. Now people know that, oh, my CEO is using it. Why don't I use it? My CEO is not afraid. My leader is not afraid. So that's where you show them, inspire them, educate them, and help them to use it.

Christopher Hutchins: 18:22

I love that because the reality is there's a lot of people inside large organizations that are taking it upon themselves because they're seeing enough, whether it's through friends, family, colleagues. They're experimenting, they're learning stuff already. So not only is it important for the CEO or the executive team to be aware of it and learning things for themselves, they also need to understand that their teams are already doing it. They may not be doing it on their job, they may be. I don't know the answer to that, but there are a lot of people that are actually learning it. And I think the mistake that we can make is really not delving into it. And sadly, this is something I'm glad you're here to talk about. For a CEO, their peer group are very high functioning people, they're very successful. Some of them may have a really great strategist that they lean on, some of them may not. One of the things that I'm really concerned about now is if there's a CEO out there that doesn't have access to a strategist, how on earth are they going to navigate? Because it's human nature. I can talk about buzzwords, I know what you're talking about to a certain extent, but I might not have enough understanding to know what are the steps I can take in the near term. And how do I avoid looking like I don't know anything? People need a trusted advisor. Maybe just talk a little bit about that. What should somebody be looking for? What's a CEO really going to benefit from and what's so important about the work that you do?

Ritu Chakrawarty: 19:59

And it's super basic. It's not starting from AI. As a CEO, you must understand where your strength lies, where are the gaps, in any leadership role. Understanding what I know and what I don't know. And then we used to call it in the old days, the Johari Window, what you don't know, what others know. So do some kind of a strategic assessment framework. Understand your operating framework. Self-awareness is the critical one. Once you understand what is my role and responsibility and what I bring to the table, now you need to know that in order for me to fulfill my role and responsibility, here is the gap. And ultimately we are humans, accepting that with self-awareness that here is the gap. Or have somebody as a strategist. Nowadays, AI tools are also helping you to at least have a starting point. Find out the basic questions at least, get some awareness. You can self-educate yourself as well. But having said that, you may have superficial level information. But you are a CEO, and you have stakeholder responsibility as well. So understanding that gap and filling it, I would say that every CEO should have an AI advisor directly reporting to them.

Christopher Hutchins: 21:42

Yes, I agree with you.

Ritu Chakrawarty: 21:44

And if you are relying on the CDO or CTO alone, I think that's a mistake.

Christopher Hutchins: 21:53

No, I agree with you.

Ritu Chakrawarty: 21:54

Because if you have an AI strategist who has great knowledge about the subject and has spent good time in the field. Again, this AI advisor should also be a mix of technology plus business. Because somebody who just knows technology and does not understand the business cannot be a good advisor.

Christopher Hutchins: 22:14

Thank you for saying that.

Ritu Chakrawarty: 22:15

Choose somebody who has delivered outcomes in a business context.

Christopher Hutchins: 22:20

That's right.

Ritu Chakrawarty: 22:20

And they now have strong knowledge of the AI subject. They should be your advisor.

Christopher Hutchins: 22:25

Your best technical people are going to have a bias towards doing technical things.

Ritu Chakrawarty: 22:28

Absolutely, you're right.

Christopher Hutchins: 22:30

And we want them to do that. But what we have to have is the two things coming together, and it is a unique person that actually can see those things and figure out how to marry them and how they should collaborate and work together.

Ritu Chakrawarty: 22:50

It's a generalist who has seen different facets of business. Has a business degree or business talent, delivered outcomes on revenue, top and bottom lines. And then they have a very good understanding of technology. That's the best combination you can have as an advisor.

Christopher Hutchins: 23:10

I totally agree with you. As we're wrapping up, if people want to get in touch with you, how do they find you?

Ritu Chakrawarty: 23:15

Yeah, best way, LinkedIn. It is very simple. My first name, last name, you can find me.

Christopher Hutchins: 23:21

Beautiful. And for those of you listening, we will definitely put her information in the show notes. Ritu, it's been amazing to be able to sit down and talk with you. I really appreciate you taking the time and being with me. I'm excited. We've got so much great content that we've come out of these meetings with, and we've still got some more to go. But I look forward to staying in touch with you. And I can't wait to have you on again. I'm sure there'll be plenty of new breaking news for next time we talk. So again, thank you so much for taking the time. Thanks for being on the Signal Room.

Ritu Chakrawarty: 23:48

Thank you so much. It was a pleasure talking to you.

Christopher Hutchins: 23:51

Thank you.