How Can AI Agents Transform the Medical Industry? | Salesforce CIO Corner

If you think about the value of AI, we have 40 years worth of clinical trial data for different devices. Right now it's sitting somewhere in a storage archive. If we can take that out, create some intelligence out of it, the way it shaped future therapies will be just amazing.

Welcome to CIO Corner. Today I'm joined by Rashmi Kumar, SVP and Global CIO at Medtronic, one of the largest medical device companies in the world that is truly transforming health outcomes through AI. Rashmi, thank you for being here with me today.

Thank you for having me.

I would start by just giving you an opportunity to tell our audience what is Medtronic and what do you guys do? What are some of the areas that you are actually making more investments in?

We work across four different therapy areas, everything to do with cardiac. Second is neurosciences. Third area is diabetes, and fourth is medical surgical, which includes our robotic surgery capabilities as well. We treat 78 million customers every year. So we say one customer every two seconds.

And in the medical space in particular, we have seen AI used for a long, long time. Now we have this whole notion about generative AI and the impact that it can have in business. Does that apply to Medtronic and are you using these types of new technologies in your space?

The way we look at AI, machine learning, generative AI across Medtronic is in two separate areas like many other companies. One is around how do we improve products, how do we improve therapies that gets deployed on the patients? But the second part of it, how do we make our company not only more efficient but also more accurate in building therapies capabilities working better as an overall organization? Since we are creating a lot of data, there's an opportunity to drive better treatment because of availability of data. So one of our device is a tiny implant which goes in your chest to do rhythm management, which is called Link. It's a device that we have which gets implanted and it captures real time data. Now the challenge was that when we get the data, it goes to the patient, the caregiver and the care provider. It has a lot of noise in it. So we have applied AI and machine learning to reduce noise in the data that comes out of it and better accuracy. Our diabetes minimed devices, which is a continuous glucose monitor and insulin pump, has algorithms built in which enables the patient to manage their diet better to keep their blood sugar under control. Our deployment of a device called GI genius during colonoscopy, because of imaging, it can actually diagnose polyp better than a GI doctor who could do it just by their eyes. On our operations side, we have recently started working on Gen AI. We had AI in machine learning and pricing and a couple other areas. But now we see our world open up. In the compliance space, in complaint handling, in sales enablement, there are myriad of areas where we are trying to apply these technologies to make our operations better and more efficient.

It's fascinating, right? Because on one end you have the ability to use these new technologies to continue to improve healthcare, to continue to improve the patients lives. And then on the other side you can use Genai to improve the business and the business processes within Medtronic, to continue to provide P and L improvements and to continue to generate value for the shareholders, which is exciting to see. You're in a place where you can have immense impact across many our employees lives better, right? Yes.

They don't have to do the work that they don't want to do and technology can do it more accurately as well as efficiently.

Now you are in a really, really regulated space. In fact, it's difficult to get to a place, maybe banking, that is more regulated than healthcare. As you implement these new sets of technologies and capabilities, whether it's to run the business or in your products to support healthcare, how are you taking into account privacy and patient rights and all things related to compliance?

We start most of our AI conversations with our AI Compass which is pulled together as a framework for us not only to be compliant, but to do the right things for our patients, employees, everybody. And it creates a set of policies around how we should be thinking about building AI capabilities, handling our patient and product data as well as supporting and operating them. Make sure we have the data access and the rights to data as well as the explainability of the results that come out of AI. One area that we apply it a lot will be applying is in the robotic surgery and one of the highest used medical apps called digital surgery, which is not only making these surgeries more accurate and training people better through the access to real time surgeries, but at the same time improves health equity and access. Because of robotic surgery, the part of the world that don't have access to many therapies will get access to these therapies. But again, as we build these capabilities, build the AI and the data capabilities around our procedures and therapies, AI Compass come first and we really set it up right from the get go.

I really like it here at Salesforce what we did is we formed an AI council and it has members from legal and it has members from compliance. From our Ethical use of technology, as well as technologists across the organization. And it's really created a good forum to create those guidelines, to create those rules of engagement as we develop and support these technologies from within the organization. Now, one of the things that's always fascinating to me is how do we make people comfortable? As I get older, I'm becoming a more frequent user of healthcare services of all kinds. But you know, now that I live in this world of technology, thinking about how, you know, these technologies come to life, I get a little more concerned about exactly how this truly is working, especially as a healthcare patient. How are you handling that and making your customers feel comfortable with the use of these technologies?

If you think about our therapies, right, we have 40 year worth of clinical trial data for different devices. Just think of the value. Right now it's sitting somewhere in a storage, archived, Right? If we can take that out, create some intelligence out of it, the way it will shape future therapies will be just amazing. But at the same time, to your point, how do we take out the people who were behind that? That's a big part of the work that we need to do. So one, in this context, a quick question around. How are you balancing privacy personalization with the development of your new products like AgentForce?

Yeah, we actually think about it in the context of building this highly efficient privacy security compliance layer in absolutely everything that happens. As we take data from our customers, we use that data to really analyze it, be able to generate insights, sites and provide outputs. This layer, this secure layer, actually ensures that one, we have an effective audit trail to ensure that we know exactly what happened from the start all the way up to the end. We make sure that we have sufficient filters and checks to avoid hallucinations and toxicity. And then we also do something that I believe is really important. And we as CIOs need to continue to keep in mind that we respect the data of our customers. And what I mean by that, Rashmi, is that your data as a customer is your data. It's not Salesforce's data. So we don't use your data to train somebody else's models and to ultimately create outputs for others based on the data that you make available to Salesforce. You asked me that question and I think about the challenges with managing data.

Yep.

I can only imagine you talked about data from clinical trials. You know, in my previous life before Salesforce, I was really familiar with HIPAA regulations because you needed to maintain the customer's data private. So you have on the product side this massive amount of data related to all of these things you do in relation to a patient. Then you have all these business data to run Medtronic. It's really interesting. How do you create a data strategy that supports managing data across those two parts of the business.

Recently we were launching our TBS device for neuromodulation and patient data. Patient consent become important. Similarly it's true for diabetes because end of the day the device goes to a patient. So really investing a lot of time in master data, understanding the data lineage as well that where did it start? I always talk about static master data which is like customer, product, supplier. But then there's transactional master data which is like order or a particular health record that comes out and really applying different discipline to those data in terms of what kind of access, security the way you talked about. Right. And then when we implement a project or an initiatives to be able to use that information to enable future therapies, it again goes through. You talked about the AI Council, we call them center of Excellence. So again it has people from regulatory, legal, hr, IT and all product areas for us to really come together and discuss as a team around this use case or this initiatives around AI. Now how do we store different data, either product data or our operational data, like many other companies, right. There are some ways we. Some places we do it very well so that it's system of records which store it and we have access to it. Some places we have real chaos and we have to pull all that together. It takes time to get to the results because we need to pull them together before we get to that work. But I think master data is at the center of it to link it from beginning to end.

I love it actually. I think it's phenomenal and I'm glad that you mentioned that because we here at salsource we are going through an MDM project as well Master Data Management project for internal use. And you know, we are a relatively young company but as you can imagine over the years our data has gotten to be many places and not always properly mastered. But you know, for our audience to hear from you in a highly complex business that if you start with mdn, you actually get a better chance of organizing this data effectively across the enterprise. I think is really helpful and we need to continue to maintain that mindset.

Kind of getting away from health care a little bit. But I've been in many companies like you and what I've seen is the most difficult activities in any large Corporate today are two areas. One is forecasting S&OP like either planning, finance, budgeting, forecasting, either areas or it's the pricing. And how do you. And these two areas actually creates a lot of customer friction if we want to create a seamless customer experience. If we fix these two areas, many companies will succeed differently. And that underpins on master data, right? The product master data and I don't know how it is in your world. You guys are much newer than us. But I can guarantee there will be issues that if you could solve, it'll be much better.

I agree. One of the areas that has been really important for us to actually use GENAI in specifically AgentForce and our agents is the whole area of customer success, customer support. And we're going all in. In fact, we just recently launched agents in help.salesforce.com so that when our customers interact with Salesforce they can get quick, very effective answers to their questions and we at the other end can simplify, can help our customer success group in the way that they manage their operation. You also work again in another very complex world where your products are highly technical. They are in the physical world as well as in the digital world. You have to have really good, solid engineers and technicians that can answer questions from hospitals, doctors, nurses as to how this equipment customers as well. How did you see the use of Gen AI in customer service? Specifically in healthcare, we have to ensure.

That we have the right complaint handling capabilities within our company to serve our patients and providers in very unique ways. We are looking at applying actually we have a couple of successful case studies in both areas. So on the customer care side, on dispute side, we have seen a lot of success. We are in a POC state, but we are trying to expand it further on the complaint handling space. There is a lot that we can do because complaints are all about knowledge bases, right? And knowledge bases are all documented. We as per regulations, have to create these manuals, insert that in these boxes which goes to the customers. But how many times you really read that, right? How do we create a very different experience when we get a complaint from our patient? And how do we collate all the data across? I cannot share all the details, but I hope you understand that there is a repetitive nature in here, right?

Absolutely. And being able to understand how many complaints of this type have happened, what has been the best possible resolutions and ultimately get to the customer in a faster, more concise and effective way to address a complaint, it's got to be huge in your business because at the end you're in the business of healthcare. Humans Supporting and helping humans. It's a huge, huge deal. So Rashmi, you know, we've been working diligently in bringing agent force to life here at Salesforce. The use of agents to support a variety of business processes, business workflows. How are you thinking within Medtronic about the use of agents to create business value?

Yeah, that's a great question. A senior tech leader came to our senior leadership meeting and they talked about how in that very large company, tech company, they're now asking everybody to think how their job can be done by an agent so that they can do higher value work. You know, in this world today, we have created jobs around ourselves. Maybe some information is in Salesforce, but I wouldn't want to go log in and look for that information. I'll ask my analyst to say, hey, can you go create a report for me where I can see how do we do bigger change management for people to be able to leverage the tech that they have, for example in Salesforce, either in service or the sales cloud. What a day in the life of a salesperson looks like. And how do we give them tools which enables them not to make a call to a call center to ask for information, but to search for it themselves. Or maybe even this agent proactively pushes that information out to whatever platform they use to run their day. Even it could be a phone call if they're overwhelmed with their email. So the way I'm thinking is and I'm enabling my team and kind of pushing them to build a strategy around how we can do a few day in life in different areas, understand what are the agents they can use and if it's a sales or a service demographic, then leverage your agents. But the point is how do we manage the change in the people's mind, in the worker's mind to be able to get used to leveraging that day in India. That's going to be the biggest hurdle, the management of change in my because.

The application of these agents can span across multiple aspects of the business. I'm not an expert in your business, but I'm imagining these engineers and scientists who are building this next generation of devices to help us, but who could actually benefit from having agents that answer questions for them quicker in relation to specifications and things that have been done in the past.

Absolutely. And software engineering as we know, Salesforce is a software as a service. Now you're maturing it to the AI and the agents. We are not a tech company. We are a health tech company. And I would like this tech to Come from outside. Where do you think this technology world is going and how we should be thinking about it?

I don't think it's practical to think that any one company is going to be able to build all this technology and then suddenly you're going to be able to bring it to life in record time and have people get great benefit from the technology. Right. I think it's going to require maintaining an ecosystem of partners that are trusted partners that will help you really build this technology in ways that you can make sure it works. It's going to become really critical that we align with the right partners, especially as it relates to building AI capabilities and solutions. That's priority for us here at Salesforce. We want to make sure that in the area of CRM, which supports healthcare as well by the way CRM is connected with customers in the healthcare space, that the CRM solutions that we make available to our customers like you are trusted solutions and that we now bring AI into those CRM solutions in ways that it's faster, it's more efficient, it's more effective for you to bring those capabilities forward for your people, your users, your business units to take advantage of and create business value. I do hope that in time the use of AI and service agents. We are big into the use of service agents to support customer service, but many other aspects of the business that that helps reduce the cost of healthcare a bit.

Absolutely.

I would assume that there will be an implication on cost. Right. Over time through the use of these technologies.

You know, we are very mission and purpose driven company. However, at the same time to protect the mission and the purpose and keep growing, our mission and purpose going from 78 million to a billion patients, as our CEO Jeff mentions many times, it does take efficiency. Yes, right. Because end of the day we are a public company. We have to give value financially. So having these technological enablements through partners like you, that's how we are going to make ourselves efficient.

Well, Rashmi, thank you again for joining me today and please check out the links in the description below and be sure to stay tuned for more episodes of CIO Corner.