CIO PODCAST: THE AI ADVANTAGE | EPISODE 3

Barbara Call 00:10

The AI era has arrived, but there's no guarantee of success. Industry estimates say anywhere from 70 to 95% of AI pilots fail to get off the ground. The question is, how can you move beyond POCs prototypes and feasibility studies to real ROI and measurable business outcomes? What does successful AI adoption look like, and what are the steps to move from potential to pay back? Welcome everyone. I'm Barbara Call, Global Director of Content Strategy at CIO.com, and this is "The AI Advantage, Navigating Risk, Reward and Real World Deployment", created in collaboration with CIO.com and Vertesia.

Barbara Call 00:55

In this episode, we'll be talking about how an AI strategy is only as strong as the organization supporting it, we'll explore the operational blueprint for mobilizing your internal technology teams, how you can help your data scientists, engineers, architects and business analysts all speak the same AI language and actively contribute to a unified, scalable platform, and how to stop working on separate projects and instead start building a synchronized it engine that drives the collective AI effort across the entire business. But first, let's introduce today's speakers. First up is Carmen Granto, CIO, Fortitude Re. Welcome Carmen, tell us a little bit about yourself.

Carmen Granto 01:43

Excited to be here and talk about AI and the adoption of this. My name is Carmen Granto, CIO for Fortitude Re. I've been with fortitude around three years now, but I've been in the insurance industry for over 25 in the technology industry my entire life. So one of the things that I've been we've been looking forward to in Fortitude Re is to advance our AI capabilities as rapidly as the marketplace is. So I look forward to talking with you today on how we address the changes within the AI landscape.

Barbara Call 02:20

And our second speaker is Ganesh Subramaniam, Head of Strategic Initiatives at Guardian. Welcome Ganesh. Tell us a little bit about yourself.

Ganesh Subramaniam 02:32

Thanks, Barbara, great to be here. I'm excited to join two forward thinking and respected leaders, Carmen and Keith, for this conversation. So a little bit about myself. I'm Ganesh Subramaniam, and I head the strategic initiatives within our digital and technology group at guardian. At guardian, I focused on developing our technology strategy, helping functions go digital first, and modernizing our governance capabilities, all with the goal of enabling the organization to responsibly scale AI across the company.

Barbara Call 03:02

And our third speaker, Keith Schlosser is a multi time CIO in the insurance industry with more than 35 years of experience. Welcome, Keith. Tell us a little bit about yourself.

Keith Schlosser 03:14

Thanks. Barbara, yeah. So Keith Schlosser, I've been in the insurance and financial services industry, as you say, for 35 plus years, currently, I'm an industry advisor for agentic AI and cyber security, and I'm excited to be with Carmen and Ganesh. We've worked together in the past, in my past life, and hold them in the highest regard.

Barbara Call 03:38

All right, nice to have you, Keith. So let's jump into our conversation. Carmen, my first question is, for you, mobilization requires not only new tools, but new skills. How do you identify the most critical AI, skill gaps within your existing IT workforce. Thanks.

Carmen Granto 03:57

Barbara, this is an evolving technology, evolving very rapidly, so our skill sets that we identified in the initial stages of AI are much different than the skills that we require today for AI. We do a gap analysis with each of the individuals within the IT organization, but also across our business, to really set the stage for what training and what programs we need within our organization in order to address AI. Skills needed within our business are very different than skills needed within the IT group, but we want that consistency across an entire organization, so everyone's speaking the same language and understands how AI will fit into our organization in the long run.

Carmen Granto 04:39

So we did the skill gap based on what we saw as the current needs within our IT organization, depending on your role within here, and developed some training associated with each different role that we see being critical to delivering our AI strategy.

Barbara Call 04:54

Excellent. Thank you. Ganesh, what are your thoughts?

Ganesh Subramaniam 04:57

Thanks. Barbara, couldn't agree more with Carmen. So just to expand on that a little bit, the framework that I typically would like to follow is: one, use surveys as a baseline. Typically, surveys help clarify workplace behaviors, confidence intervals, bottlenecks and readiness for aI don't rely on assumptions. Next, look at task and role analysis. As Carmen said, this is an evolving technology. How does AI shift specific tasks, model evaluation, secure development, data wrangling, decision making. Then you got to take these skills and map it backward to the tasks. Third one, capability and platform audit. You got to benchmark your current capability to execute your life cycle, make sure you have governance in place. And then, of course, there are telemetry and community signals, which help you identify from the systems, from your architecture guilds to surface real world friction points and your capabilities. In a nutshell, you typically want to identify the four priority gaps, ops and automation skills, Product Engineering, secure, AI and AI, augmented, workflow, proficiency,

Barbara Call 06:10

Perfect. Thank you. My second question and Carmen, I'll start with you again. What was the most effective strategy training, recruiting or strategic external partnerships that you've implemented to quickly get your legacy IT talent ready to support and manage a modern, scaled AI engine.

Carmen Granto 06:31

Thanks, Barbara, we use all three of those strategies that you mentioned. I briefly talked about our internal training program and customizing that to the individual and the roles within our organization, and that has been what I'll say, the bedrock of our strategy around building our talent. They bring not only the talent, the knowledge of our data, the knowledge of our environment, the knowledge of our company, and then acquiring the new training through the training, acquiring the new skills to be able to deliver on AI. But we've also brought in some external strategic partners to jump start that and to really bring some thought leadership into our organization that we didn't want to wait on the training program to develop within the house. Those strategic partners have what I'll call jump started our progress into AI. We've also adjusted our recruiting, and while we may not have many positions right now that are specific to AI, we have adjusted our recruiting to be more focused on AI capabilities as we start bringing in different roles within our organization. So those three things, as we change through the process, will have been able to advance our AI technology, not only within it, I should stress. But also within our business, our data practice is very involved with our AI and our AI strategy, and they employ a lot of our similar training, plus the recruiting and their external partners as we share this. So it's not just technology, but our business as a whole is we believe everything needs to come together in order to produce AI at scale and to be able to leverage those AIs and take them from a POC or prototype and into a production process.

Barbara Call 08:07

All right, very interesting. Ganesh, what are your thoughts?

Ganesh Subramaniam 08:12

Yeah, so insurance talent typically understand their domain very deeply. Are you talking about underwriting, claims, risk assessment? So as you look at your skill sets, the strategy should really be to add AI capabilities to that foundation, not replacing that domain expertise. So companies are committed to reshaping workflows, see higher employee support and engagement. And then, as Carmen mentioned, you can bring in partners to help you accelerate you can do strategic hiring for some senior machine learning operations capability product leaders who can elevate the entire teams through expertise and leadership. And then, of course, you have specific task based training, as we've mentioned earlier, where you really focus on their tasks that people do on a day to day basis, and target it to that and not generic training. So I believe a combination of those three would really help elevate the level of training and skill sets in the organization.

Barbara Call 09:07

Great, nice ideas. That's excellent. So my third question here, What is your advice for how CIO and other business leaders can help the various team members speak the same AI language and actively contribute to a unified, scalable platform. Ganesh, let's start with you this time.

Ganesh Subramaniam 09:28

Right. So, as you mentioned, Barbara, the key is to create a shared AI operating model. And that's, of course, consists of common vocabulary a unified life cycle. And really driving that alignment starts with, starts with the following. A shared lexicon. You need to have a lexicon that defines fairness, robustness, bias, risk levels. All of these are very, very specific to AI. Secondly, you got to create cross functional AI pods. Want to break down silos, because once you have silos, the resistance and the and the cross functional missteps in terms of languages and miscommunications can occur. So it's important to drive cross functional pods and then start to create governance and platform level guard rails so that the organization can intelligently and very, very deliberately leverage enterprise capabilities with consistency and scale.

Barbara Call 10:30

Carmen, what are your thoughts?

Carmen Granto 10:32

Ganesh, I know it's been a few years since I've worked with you, but obviously that was an excellent answer, and no wonder why I liked working with you back in the day so much. I agree with Ganesh. Breaking down those barriers, understanding that this is not solely an IT initiative when you're talking AI. A lot of technology, and I see my role as a CIO is to bring the right technology solutions to play, to solve our business problems, or to solve our business demands and goals. However, breaking down the barriers, understanding the silos and having a unified approach to AI is what we're really focused on. A lot of our AI initiatives, a lot of the use cases are not generated from within it. We enable those use cases. We enable those solutions to be put in place and apply the right technology. So while it plays a key component in how do we apply the technology, and what technology to apply, and how do we get to that scalable platform, we work hand in hand with our business. So we have steering groups. We have different methods that we use different pods, different ways that we bring teams together that focus on those goals across not just it, but across our business, and I think that is what I would say was be the key for us is really ensuring that we have the right resources and the right skill sets involved to deliver on what we're trying to do for AI,

Barbara Call 11:53

Okay, great, Keith, love to hear what you have to say.

Keith Schlosser 11:58

Yeah, I'll take it down just a little lower level here. So I would say the first thing that a CIO should do with his business or her business leadership is make sure that everyone understands what AI is in the context of the company. Not everybody understands it. Not everyone understands the difference between generative AI and agentic AI. And so as we start thinking about bringing everyone together a common AI language for the company, I think it starts with making sure everybody has the basics down. And I think we need to provide some education. We need to work at demystifying it. Then the next thing I think is really important is many folks are afraid of AI. They've heard about it in the news. They've seen some of the headlines that maybe don't make them feel that confident. And I think every company has an obligation and an opportunity to help people understand that AI can really be important to the organization and, more importantly, important to their career. Important to their careers. And I think that if they'll spend the time doing some of that foundational communication and learning and providing tools for people to really get a handle on what it is, how it fits in the context of the organization, and how it can help them. That's that's helpful, right? That that will start to get everybody on the same page. The other thing I think, is really important, and the guys already kind of talked about it, but what does success look like for AI within the organization? So many times, CIO and business leaders start projects and they fail to define what we're trying to do and how we're going to measure it. And how do we check to see and hold ourselves accountable to what success is, to what the ROI is to, why it's good for the customer, why it's good for the company, why it's good for them. And then I would say the last thing that I'd add is, with any new technology, I think Ganesh and Carmen would agree, we often see a lot of skunk works in pet projects, and this is not something that I think is helpful in the context of a genetic AI. I think we need to share that there is a platform, there is a strategy. We do have a concept of the tools and the projects that we want to go through, and as they both talked about with governance and and guardrails, set those up, make sure that people understand them and and when everybody is on the same page at those foundational levels, I think your chance of success is pretty high.

Barbara Call 15:22

All right, Carmen, my next question, I'd like to start with you, where should CIOs and IT leaders start to build a synchronized it engine that drives the collective AI effort across the entire business, and that's versus working on separate projects?

Carmen Granto 15:39

I think we start this with a discussion around where we've been with with AI at least within four to two, and I think other companies would be similar to this. We've introduced multiple AI solutions into our environment, whether that be some of the some of the big GenAI tools within our environment, and some of the agentic tools that we're doing. We also have a lot of vendors now building AI into their products and into their services that they provide us. So when we look at where do we want to bring that together, it really is in the user experience. On the GenAI side, do I need to know which tool to go to or which engine to go to in order to ask certain questions versus other questions? And so we are looking to build what I'll call a more universal GenAI solution within our environment that takes the model away from the user decision and really allows them to just ask the questions that they need to ask, prompt the tool correctly, and then we'll decide and drive that to the GenAI engine that best suits that. We're also on the agentic side evaluating now, how can we kind of move through the clutter and decide on what is our agentic strategy? We do have multiple tools in our environment today with a desire to streamline that and leverage that a little bit more. That's more from the technology standpoint, What technologies do we want to bring in use? But it also talks to how do we approach that, and what is our vendor strategy? How does that work with what we're delivering built internally. We have a few platforms that we're developing within house now, more as prototypes and PLCs to see how we want to go about doing that. I'll say again, that is a little bit more focused on the technology. I am the CIO, so I'm very concerned, very involved with the technology of that. But we're also talking about how and where do we start bringing that together from a business standpoint. Where's our steering committees? Where is our guidance around what technologies, how do we want to employ those? What's the experience we want our employees to have, our users to have? And we also are bringing that together to make the decisions, not just from a solely it and solely technical basis, but from an experience and a productivity type approach as well.

Barbara Call 17:56

Yeah, that's great advice. I like it a lot. Ganesh, what are your thoughts?

Ganesh Subramaniam 18:02

Keith talked a bunch about, sort of some considerations, on, on, on, sort of making sure that there's lack of skunk works, ensuring the right guardrails. Carmen was spot on, a very comprehensive answer. Essentially, what I would say is you start with your operating model foundations. What we should be focused on is to really reshape key functions, rather than small scale productivity focus initiatives. So specifically, the pillars that I would look at is creating unified AI platform, make sure that that's reusable, provide the ability to create components that can be leveraged across lines of business, and which helps drive AI's long term value. Again, we talked about pods change ways of working, create cross functional AI pods with shared accountability, so that there's a real push towards ensuring that AI is being leveraged to solve the big rocks of the business problems. And then when you look at establishing governance, you have to implement a risk tiered model review AI is a different beast when it comes to governance. So this has to be done with a lot of care and deliberation. And it's not exactly the same as what you would do for any other technologies, because the model risks are very different from use case to use case, and when you look at AI platforms. And then finally, as Keith had mentioned very eloquently, that you've got to look at a value based prioritization. You cannot go and do one off projects, and you can't do a lot of skunk works. Otherwise, we'll simply dilute the focus on the value that you're going to drive with AI.

Barbara Call 19:52

Thank you, Ganesh. Keith, what are your thoughts?

Keith Schlosser 19:55

So to add on to what Carmen and Ganesh have said. I think there's a couple of things that are really important here. One, I think, as an organization, as leaders, both on the business side and the IT side, we have to ask ourselves a question, do we want point solutions and solve a couple of problems because we're under pressure from the C-suite and the board to show progress? Or do we want to start thinking differently about how AI can solve problems? What I mean by that is oftentimes, in IT development, we tend to solve a specific problem and then hand that problem off to the next process, and there's some sort of data store behind the scenes. And code is written, and it grabs the answer, and it goes on, and it starts the next process. And the next process. I think selecting the right AI platform allows us to start joining traditionally disjointed processes, and and once we get a handle on that concept and we understand the importance of the data to the overarching value of an AI effort, I think that's where we start bringing things together, and that's where The value of of a platform versus solving a specific set of issues or challenges that lie before you. That's where that's brought out. Again, talk about guardrails and governance and security, all of the stuff that we've all mentioned so far, really important. But I think it starts with understanding, what are you trying to achieve? Do you want to make the board feel happy and excited that we're making progress? Sure. Do we want to do that at the expense of picking the right platform and looking at the value that AI can bring, looking at that a bit differently and joining up disjointed processes? I think that's a question that needs to be asked by the leadership team.

Barbara Call 22:04

All right, thank you, Keith. And so my last question for today, looking forward, what's your prediction for the next big change in AI, and how can it and business leaders prepare? Ganesh, I'd like to start with you.

Ganesh Subramaniam 22:20

Okay, well, that's that's an interesting question. So as I said, Here, we look at all the work that we've done in AI in the industry. The area that I see that's a very likely outcome is, is AI evolving not just beyond a pure agentic solution, but really multi step systems that orchestrate actions across tools and data, and we're already starting to see some of that in the industry, where they're starting to create create a few multi step processes, such as interacting with customers, processing transactions, coordinating follow up actions. But when you start to take that a step further, it really starts to get into observable policy, where agents, they're embedded in redesigned workflow. And it's really representing the next evolution, and it's very, very important for companies to be able to master that transition and help disrupt the industry and define it for the next decade.

Barbara Call 23:24

All right, thank you. And Carmen, wrap us up.

Carmen Granto 23:27

Yeah, Keith, broached that with the multi task mentality, and Ganesh, I fully agree with your, your assessment there, I think it's going to grow. I think to date, it's been very task focused. How to automate a task, or how to apply AI into that task. And once you start getting across the multi task or the value chain of a process, I think you're going to see that a lot more get applied there. The other thing, which I think I'd mentioned, that is not maybe specific to AI, but with the with the billions, and maybe we're into the trillions now with investments I see companies making around the world in an AI technology, I think we have to get a little bit used to the to the to the speed at which this is being introduced. New model comes out on a very, very regular basis. The speed that we can integrate that and jump on to that and understand the impacts of any changes with a new model, with switching from one model to another, which is what we want to be able to do in the future, being able to have a well oiled process that we can do that and make sure that we're comfortable with that. And then the last thing I'd say there is, folding in our vendor strategy with that as well. Every vendor that we work with is introducing AI, and it's debatable what is, what is truly AI within their tools and what is just a rebranding. So figuring out and aligning our strategy and our approach with theirs as well, we rely how. On certain key vendors within our organization, and making sure that our strategy and theirs align. When are we going to use theirs with when we can replace it with ours? Is going to be something that is going to move at a pace that I don't think we've seen before. So I would say those are the areas that I'm preparing for for the future.