The Deep View: Conversations

Priya Srinivasan: We've seen a lot of digital transformations. We've seen transformations or we've heard of transformations even before going into the previous century. This feels very different. This transformation is at a different scale and pace than anything before. The value is still the same. It's about making people's lives better. It's about making for enterprises, it's about making clients' lives better. So the value, the core aspect remains the same. But I would say the pace at which things are changing is probably one of a kind. And it requires a lot more adapting than we've had to in the past.

Sabrina Ortiz: Thank you so much for joining us today. We have an exciting announcement today from you and your team. But first, I would like to get started by you just telling me a bit about yourself. What do you do here at IBM now? How did you get involved in this crazy AI space? And what are you working on now in particular too?

Priya Srinivasan: First of all, thank you for being here and having this conversation. So my name is Priya Srinivasan. I've worked in IBM for the last 25 years almost. I've been in IBM software the entire time. I've worked across multiple products that have done a lot of different roles, including professional services, product development, product support, center of excellence, a lot of different stuff. And I worked across multiple, multiple products from data to AI to automation. And I've just experienced it over the last couple of decades.

Sabrina Ortiz: 25 years, I'm sure you've seen almost every digital transformation that you could imagine. How does that compare to this moment of AI that we're living through right now?

Priya Srinivasan: So this question comes up quite a lot, even in our discussions of, yeah, we've seen a lot of digital transformations. We've seen transformations or we've heard of transformations even before going into the previous century. This feels very different. This transformation is at a different scale and pace than anything before. Now, the value is still the same. It's about making people's lives better. It's about making for enterprises. It's about making clients' lives better. So the value, the core aspect remains the same, but I would say the pace at which things are changing is probably one of a kind. And it requires a lot more adapting than we've had to in the past.

Sabrina Ortiz: Have you applied any of the same, I guess, mindset or skills from trying to adapt to the other transformations now when you're thinking about deploying products and services relating to the AI space now?

Priya Srinivasan: The core thing, it always comes down to the right culture, right people, right skills, and then the right tools irrespective of the transformation. So having that adaptable mind to learn, be curious, is super critical. And I would say that has always stayed and we're continuing to apply the same thing, really understanding the space, understanding what is it the problem that we're solving for? It always starts with a problem. or an opportunity. What are you actually trying to address? So I would say that trend continues. The culture and people are a very, very key part. You always take your people in the journey. You always transform the culture along with the tools that you bring in and then finally use the best tool for the job.

Sabrina Ortiz: When you're identifying the problems that you want to solve, I'm assuming you're talking to clients. Is that also how you, I guess, first find what you need to solve before you even start building that product?

Priya Srinivasan: That's exactly how we're getting to that point, talking to clients, talking to our partners, having those conversations to really understand what is top of their mind? What are the places that is opportunity for them to make it better or problems that they are trying to address? And then ideating over that and making sure that we're really addressing that and what would that translate into as use cases and what type of technologies would really solve for that use case? So yeah, start with the client, always. Start with the problem that you're trying to address.

Sabrina Ortiz: While it feels like IBM has been in the space forever, right now there are key players in the AI space across the stack. And I'm curious how you got involved specifically with what the announcement we're here to talk about today, Sovereign Core and what you heard from clients that fueled the birth of this product?

Priya Srinivasan: I'm very tempted to say my boss said so. But putting that aside, so if you really look at, and I have to explain how we've looked at Sovereign Core, so you have to bear with me on the long explanation to answer that question.

Sabrina Ortiz: That's perfect.

Priya Srinivasan: So the way we approach Sovereign Core is, digital sovereignty is becoming more and more real and the world of AI that we're living in, it's accelerating that further.

Sabrina Ortiz: And before you even keep going, let's just define for our audience too real quick, digital sovereignty, what do you mean?

Priya Srinivasan: So it's the whole transformation of how do we bring that digital experience, right? Where at the end of the day, everybody's trying to do a few things. They're trying to increase that client experience, make it better, drive more revenue, save cost, or modernize the systems. It all comes down to a few things that everybody's trying to do. And doing it in a very highly digital experience and the whole digital modernization is not a new thing, but just doing it in a highly sovereign manner. And I want to break it down what that actually means. Forever we've been talking about data, like making sure, hey, the data is where the data resides, what the location of the data is well understood. Then from there, it has gone into other aspects of, hey, who actually operates my system? And how much can I rely on the technology for my business, right? For my business to stay resilient. And from there, in the world of AI, is where are my models running? Is my inference really governed? It's expanded to a lot of things to truly make it digital sovereign. Now, your question on how did we begin about all of this? So we work with a lot of regulated industries. We work with governments, we work with regulated industries, we work with local cities, countries, that is the business IBM has been in across, you're talking about countries around the world globally. And we have seen these needs over a period of time, but more accelerated now that there are a lot more, I'm using the word constraint, but it's not constraint, it's compliance. Maybe there's a lot more compliance that you need to follow certain frameworks and rules of operation.

Sabrina Ortiz: Internal or external, you would say both.

Priya Srinivasan: External and internal, both of those. Now, speed is very important. Speed of innovation is very important. But at the same time, you have to do that in a highly compliant manner for these industries. So talking to these clients, like back to governments, cities, countries, regulated industries, we are seeing more and more the need that they are expected to be highly compliant to needs of either external regulators, like you asked, or internal company policies and rules and regulations. And that in the world of AI just became 2x, 5x, 10x around, hey, am I, where's the brain? Where is it really governed? Where is the model running? Who really has access to it? All of those questions have become more prevalent now. That's how the inception of this began. All these threads that we're seeing around the world, okay, how do we address that? And how do we address that very holistically?

Sabrina Ortiz: And I'm assuming with the now widespread adoption of agents that's only more important because now they're actually taking action, people say, and people want to understand like, what are these things? Where do they come from? What can they have access to and all of that?

Priya Srinivasan: Agree. So building agents itself, I think people can do it. And that is going to be a proliferation of agents everywhere and agents are gonna start executing work, right? We've been saying AI is no longer just the assistant, AI is also gonna do some of the execution of the work. Now it's super important what becomes difficult is not necessarily understanding the AI regulations, not necessarily building all of this, but what makes it hard to take something as a POC into production is really the compliance, the risk aspect, the governance behind it. That's the real crux of the matter. And addressing that in the world of agents, doing it in a very trusted manner is where Sovereign Core plays its role.

Sabrina Ortiz: Yeah, so can you tell me a bit more about how it actually, Sovereign Core takes place or actually works for clients who might be hearing about or people who might be curious about how it would actually look in their everyday workflows or lives?

Priya Srinivasan: The big question that I get asked all the time is, what is Sovereign Core? Right, it's software. To start with the simplest is, it's software that IBM provides. Now this software, it's essentially the way we built it is it gives you the foundation technologies to quickly stand up what you need to start building either traditional workloads or AI based workloads. And the simplest way is the control plane of everything that you need to do from your keys, your access, your secrets, all of that is all managed within your boundary. So A, the software can quickly stand up a foundation for you. B, when it stands up, it makes sure the control plane is within the boundary. C, it also makes sure it's secure, right? It's everything is inside your environment and you can deploy it anywhere, by the way. You can deploy it on-prem, you can deploy it in an air-gapped environment, you can deploy it in cloud, again, it's software so it gives you the options to deploy anywhere. And we have built it with the principle of we want to be hardware agnostic, infrastructure agnostic so that clients can leverage their existing investments and build on it. And then, okay, you've done that. You've stood up this environment, great. It's highly compliant, it's all within Sovereign. Now what? Right, like that is the value part. It's really towards two things. One, being able to bring AI down to hours and days. If you need to take an application, take it all the way from POC to production, you're talking months, like months of effort, but how do we bring that value down into hours and days is a core principle behind building this environment. And what really prevents us, anybody can build an AI application over a weekend, but what prevents them from really taking to production is things that I mentioned around compliance and risk and all of that. So how do we accelerate that for them to be up and running quickly with AI? That's one part. Second, Sovereign Core comes with this extensible catalog. We have always believed that you should give choices. It could be IBM technologies that we have, it could be open source technology, we work with an ecosystem of partners, it could be anything that a partner ecosystem can put in. So we have this extensible catalog that gives clients the option to pull down. So without saying it, I'm kind of I guess saying it, it's like cloud in a box or in a software that we want to ship and then you're able to get into this extensible catalog from there.

Sabrina Ortiz: Now it's time for a word from this week's sponsor, Microsoft. What does it take to go from an AI idea to your first paying customer? Most teams are experimenting with AI, but experimentation doesn't generate revenue. Shipping does. Microsoft AI Envisioning Day is a free video series built for developers and software companies ready to turn ideas into real products. You'll learn how to take an AI idea to a working MVP that customers will pay for using practical patterns and guidance so that you can build something that actually generates revenue. No fluff, just clear frameworks and steps you can start using right away. If you're building with AI and wanna start closing deals, not just shipping demos, this is how you get there. You can get started today at aka.ms slash The Deep View. That's aka.ms slash The Deep View. And we thank Microsoft for their support of The Deep View. And now back to the show. I have a question. So this is trying to solve a really important issue of making sure you understand as an organization, as if you work, whether in IT, and you're actually setting up all the access, or if you are actually just a team leader and wanna make sure that all of those protections are in place, this is ensuring that a way, an easier way to make sure that that's there. But that requires a demand from the actual client and an awareness that that type of, I guess, sandboxing or that those protections, those governance, that all of that is necessary and important. Are you finding that client are aware of the importance and the risks involved if you don't have a layer of protection in that way?

Priya Srinivasan: Absolutely. And clients are trying to solve this. Enterprises are trying to solve this, but the way they're doing it is using a lot of policies and a lot of control. Could you do this in a much more manual way, in a much more manual manner? Sure, with more workflows, with more process, with a bag of parts, absolutely you could do it. But what we wanna do is really do it in a cost effective manner with an experience that is very integrated. So the actual need to do it is understood, but the way it is done is probably done through a lot of manual audits and workflows and controls and process. And that's where we're saying, well, let's do an automated compliance. Let's do it in a way, and by the way, it's not a one-time thing. Meaning you don't set it up once and I'm done, right? It's ongoing and we've taken the principle of how do we do ongoing compliance and how do we prove it to you? It's really sovereignty with receipts, right? Which means with proof and we will give the proof through our compliance center. So to answer your question, yes, it's needed, it can be done and it is done through a lot of manual ways. We just wanna do it in an ongoing automated manner with proof and evidence right in the box. And that's the design principle with which we built this solution.

Sabrina Ortiz: How do you also balance having all of the right, again, limitations or access and all of that with not slowing down the processes that people, that I'm sure your clients want, like the speed, the efficiency, the optimization, how do you balance the two to make sure that there's, you know, the right safety precautions are in place, but also that it's not slowing down what the goals are?

Priya Srinivasan: It's probably one of the hardest problems to solve. Actually, I would add one more mix into it, cost, right?

Sabrina Ortiz: Right, of course.

Priya Srinivasan: And everybody wants everything, right? We want speed, we want it to be highly cost-effective and we want it to be compliant, right?

Sabrina Ortiz: It's a balance.

Priya Srinivasan: Yeah, there is not the perfect answer, you know, yes, always you're gonna meet everything immediately day one. It is definitely a balance, but we have taken an approach that clients have to worry less about the compliance part and more be focused around the innovation aspect by bringing a lot of this principle into the design of the solution itself, like I mentioned. And we definitely wanna do it in a very cost-effective manner for them, right? So to answer the question around speed of innovation, we have tried to put the right solutions in the box as a base service versus give options through the catalog. That way, the ones that you want absolutely, that we believe you need to be up and running day one, day two, you have it out of the box as base service, out of the box as agents. While, yes, there is going to be others for you to have a bigger ecosystem of story, you are able to pull it down. So we've tried to balance it that way and sovereignty from day one really matters. Meaning it was not an afterthought in our design. We didn't build the product first and then think where do we need to plug it so that we get, now we said day one, when we design the product, what is the core value? We called it Sovereign Core for a reason, right? And so how do we build that from our first session and discussion around the architecture to today as we GA the product? We have ensured that that's through the thread of the design so that we can balance speed of innovation towards the compliance versus and as well, do it in a very cost effective manner.

Sabrina Ortiz: Yeah, I would also like to touch on or I guess double click where the risks of not having taking the right precautions to having a compliance in place.

Priya Srinivasan: I use a very extreme example, but that's just an extreme. There was recently news about a database getting just deleted by an agent. It's an extreme example, but it's a real example. Right, things can happen, right? Things can happen where you don't know who has access to things, the data that it's using, right? Is the answer even correct? We've all used everybody, I think one thing that is familiar is some form of chat interface. Everybody is used to getting some response back. And we all know sometimes the answer we get is not the right answer, right? It's okay, we're using it for personal reasons like what's the best restaurant in town, it's okay. If the answer is wrong, you can't do the same thing for a banking loan. You can't do the same thing if you're recruiting someone. The accuracy really matters. And on what information it is using to respond back, that information is very important. You need to be, it needs to be very transparent.

Sabrina Ortiz: Right, that's the only thing.

Priya Srinivasan: Yes, and so it needs to be very auditable. It needs to be highly observable, right? All of those is key when you're running it for real enterprise needs. And if you don't do that, the inaccuracies and the examples that are said, and especially as it takes actions, we live in a very rule-based engine world where decisions were based on rules. Those rules were created by humans. We said, if this, do this, if that, do that. So it was, we created it. So we still had, it was our brain.

Sabrina Ortiz: Control, right?

Priya Srinivasan: Control, I love the word. You said the word that I would have loved to say control. We had the control. You're kind of giving the control away now. When you get that control away to something, you do not know the ramifications of the actions that can take, and my intent is not to scare. My intent is more to, it's a real thing.

Sabrina Ortiz: It's a real thing, of course.

Priya Srinivasan: You have to be cautious. You have to know the tools you're using. You have to know where your models are running. You have to know your inference is governed. You have to know all of it so that unexpected things don't happen. And that's really why there's a stat that says, there's a 80% or so POCs, they fail. They don't go to production. You have to ask why? If building AI is easy, why is going to production that difficult? Because you lose it in the rest of the governance aspect. And that's why it's very critical so that you can still run a highly safe, secure, trusted enterprise.

Sabrina Ortiz: How do you factor in things that are just out of any of our control? Like for example, AI models are bound to hallucinate because they're not quite really synthesizing or understanding information the way you and me do and the way we think and reason, even though we use the term reasoning all the time with describing these models. What they're really doing is analyzing, the information that we're trained on, information that they're pulling on, and kind of using that pattern to predict. So that obviously leads to hallucinations. There's not really anything yet that has been done that could significantly reduce, that's just the matter of the technology, is how technology works. Then how do you factor that? Because that's out of your control. Like what these models can do. And then also, I guess, give companies a reassurance that, hey, with our offering or our product, we could better gauge that accuracy or we could better gauge, reduce the risk of those who say just like to your point, banking or medical care or any of these industry use cases, you can't afford to have those hallucinations.

Priya Srinivasan: Fair and a very important question. For years, we used to say security by design, yes, security by design is super important. So is governance by design.

Sabrina Ortiz: Right.

Priya Srinivasan: And governance does not mean more meetings, more process, that's not at all a governance. It is completely taught through and being able to observe things every step of the way. That observability really matters.

Sabrina Ortiz: And that includes what like tagging, like making sure you know, like by observability, I mean knowing exactly what model was used, who triggered the action. All right.

Priya Srinivasan: Correct. Who triggered the action? What was the response? Being able to trace it back is super key. Now, we do with our solution, give the choice of bring your own model. So we're providing you the option, whatever model you can bring. But back to the governance by design, if we have the right governance and not trying to sell our governance solution here, but we do have governance solution that does exactly that, right? Where we are able to orchestrate these agents in a way, it's not just about building, it's not just about managing, it's not about just running, but it's also about observing them. It's also about auditing them. It's also making sure that you're running it in a very, very governed manner. So that approach is not a day 30. It's a day zero thinking. And it has to be in the design of it to really prevent back to what you were saying around hallucinations, prevent some of the ramifications as a result. guardrails, more and more solutions are putting the right type of guardrails in and to make sure you have that. So your answer is within the reasonable acceptance as well.

Sabrina Ortiz: With what you were talking about observability, a lot of it is understanding who triggered what agent or what entity actually did the transaction or like in the extreme case you mentioned, deleted the entire inbox or the whole platform. How do you, again, things out of your control? Like for example, if employees don't have the right amount of AI training or upskilling, you might be able to trace it back to them, but are they really at fault if they were entrusted to use an agent without really understanding ramifications? That's not quite something that you can control. Similarly, what if you traced it back and the agent decided to go a bit rogue and then it's like, is it the fault of the person who triggered it or is it the fault of the agent? How do you, again, more philosophical, harder to grasp questions? How do you tackle those when you're also, again, working with clients and helping them mitigate these risks?

Priya Srinivasan: That's what is very different. So one of the things I do in IBM is I run software support and support for IBM software. And we use a lot of AI in our client engagements. We use it before a client even opens a case with us, like we call it deflection, so making sure that they're getting the help on the system before they come. Once they come in, then we look at how to give them automated responses. But if that's not satisfactory enough, then how do we speed up the resolution process by leveraging AI? So we have a three-pronged approach. We've been very thoughtful because you're giving an automated response to find trust, it's IBM. Yes, it says generated by AI, but it is generated by AI from IBM. There is a trust element. So the amount of, not just the testing, yes, we did the testing, the amount of slow rollout and the careful validation. We did all of that over a period of time, but we did deploy trust but verify. Ever underestimate the importance of the people and the expertise and the domain knowledge and all of that? That is the biggest, biggest strength we all have. And our engineers feeling comfortable, right? And they were the ones who built it. Our support engineers built this for our clients. So them being part of it and them feeling like validating everything, saying, hey, I feel comfortable with the model, with the model's response, with the accuracy of the response, with the way it's responding, because the client's gonna see it even before we see it. Right? That requires a level of trust but verify and that trust but verify is by the experts. And your experts are never out of the loop. They're like the most important things, right? Like into this whole system, right? We're talking about technology, but there is no technology without people. So our experts are at the heart of it and they are there to make sure it's validated and we'll have a level of comfort before we're turning this to the clients. And that's how we've taken an approach and we've turned it over to clients in so far. It looks so good. I would say that for any use case and AI has its strength in some of the use cases, customer support is one that I mentioned. HR is a great place, procurement is a great place, sales is a great place, marketing, these are all where areas AI has its strength but it is that validated approach and making sure our domain expertise of our people is in that system remains still key. And in some places, I would say, maybe we've taken a safer approach where the approval is done by a human, the work is done by AI, but the blessing is done by a human before it goes. In some places we've been a lot more liberal with the AI does the work more with the human watching it.

Sabrina Ortiz: IBM has also taken the approach and I think you touched on this before where it is, a lot of its offerings exist to support IBM tools that are other IBM offerings, but also third party solutions too. And I think that's done to make sure that clients have choice, to make sure that they can implement IBM solutions no matter what their workflows are currently looking like and anything of that sort. But with that, it also opens the fact that there's some right risk or lack of control, going back to our word of earlier, that you don't have the entire, you don't have complete control over knowing how a tool was trained, the risk in my involved, all of those type of things. So how do you mitigate that in itself too?

Priya Srinivasan: So with AI we've taken like three pronged or multi pronged approach. Of course we have AI solutions that we sell.

Sabrina Ortiz: Right.

Priya Srinivasan: And of course we use AI internally and we're big proponents of client zero and you're truly driving that productivity. That's number two. There is another angle. We have been building products and solving enterprise needs for years, for decades, over a century. Bringing value into those workflows, those systems of record using AI has been a key part of how we've been approaching. If let's say I'm gonna take any one of our products as an example, right? Like just say a DB2 or a Cognos or an MQ, these are all products that are deep in customer assistance. We're infusing AI, but these products integrate with other tools, other services, other models. We let clients choose their inferencing engine. We let clients choose the models. We don't want that to mess anything up. You know, technology that is deep into their system. So if you're using, let's just say, DB2 as your backend database, or you're using Cognos for your reporting engine, or you're using MQ for messaging, you're using any of this and you're deep into your systems. And these are integrated with other tools, other models, other services and all of that. We wanna make sure we're still protected in what DB2, it's a highly secure database. It's a highly resilient, highly secure database. So we have enough guardrails in place around the actions, around the information to make sure it's still the core value prop of DB2 being this highly secure, highly performant, highly resilient database is not something that is disrupted. So you do have to build those solutions when you live in a world where tools talk to tools, agents talk to agents services, and then as we all know, there's standard protocols, standard frameworks, standard ways these interact to make sure that it all stays as much as possible, governed and risk-free.

Sabrina Ortiz: You mentioned having a client-zero approach. I find pretty fascinating, and I love it when companies do this because if you're going to put out a product in the world, you might as well also be willing to use it. Were you always comfortable implementing AI into your workflows or into your, I guess, everyday thought processes? And have you found the same openness and willingness from your teams, from your colleagues?

Priya Srinivasan: All your questions are good questions, but this is a very good question. In all transparency, it came from our chairman. He declared that as an objective, and this was several years ago. And it was very simple. It was simple by the fact that we're a large company. We have systems that were born years ago. We have systems that are modern, cloud-native. So we have a combination of different types of systems. We use different tools. People are large employee base. We work across all the industries and across so many different countries. So we're big enough that if the product is fully scaled for our environment, we have a winning story to go and tell our clients. So it was a very conscious decision he made, and he said, we need to take client zero seriously. To do two things. One, yes, help with our productivity. And we started with our HR use case. If my memory serves me right, it was a very first use case. And I'll get to your core question on believing in it day one versus believing in it later. From there, we've expanded across every team in IBM. But it was a conscious decision on not just on the productivity part. It was also getting the benefit of the scale to make our technologies highly secure, highly scalable, highly resilient, all the good stuff. And it drives that value. Now day one, did everybody jump up to it? Probably not. But they saw the value. And I'll use HR as an example. Because if you look at managers, they do a lot of great work. They work with clients. They work with the team. They work with the product. They're also doing the operational stuff. And I'll give you a personal example. As a manager, pre-AI, if I had to make somebody a manager, I'll have to follow the list of instructions, step one, step two, step three, step four, step five, and actually go and click. I have the instructions on one side and then click and do it. Now what we have is something called Ask HR. You go to it and say, make whatever, whatever, as a manager. It'll listen to you. Are you sure? Back to your question on making sure. AI doesn't go wrong. It says, and once you say yes, it says submitted. You want a report generated? I used to actually ask a person to help me to generate a report. Can you get me this data? I want to see this information. Now, those report generations are all self-service. Now that we saw the value, we saw the benefit, now the belief in the system has expanded. So it started top down. But now we've all seen the value. We've seen the value in two fronts, our own productivity. And now we're able to talk to our clients with a lot more confidence. We know this will scale this way. We know this is how it needs to be done. Now it's in the culture. That's why I always believe it's always the answer. It's tool, people, culture, process. Everything should coexist for a successful outcome. And I would be proud to say we've figured that angle pretty well.

Sabrina Ortiz: In your personal everyday work flows, I'm curious, how much you reach for AI? Do you use AI tools? And what are your favorite use cases of so?

Priya Srinivasan: I use AI quite a bit. But here's the thing, I don't think I use it to the full capacity I should. OK. I think I could use more.

Sabrina Ortiz: And why is that? Why do you think you don't use it as much?

Priya Srinivasan: I think it is just the comfort of doing things sometimes, the way you're used to doing it. And I am somebody who believes in it. And even I fall to my way of doing it rather than leveraging the stuff. So it's a conscious shift in decision I make. No, I can do this faster and better if I used AI.

Sabrina Ortiz: Right. And the thought process is sometimes one of the biggest challenges, right? Like getting yourself to be like, oh, wait, I actually can use AI for this thing. Like, bye. And it's that constant reminder. And then it gets into your DNA and you're able to do it.

Priya Srinivasan: So yes, I use it, but I should do more and more and more. That's how I'm going to create more bandwidth and capacity for myself. Now, the place that there are a lot of places where I absolutely love it. But personally, the place that I've loved the most is, I mean, there's a technology called Bob and it's a development partner. And you can very quickly stand up information in a way using Bob. I used to be spreadsheets and PowerPoint and oh my gosh, like death by a thousand paper cuts. Sometimes I feel like, oh, death by so many spreadsheets. And that we've kind of really, really gone away. Even though we have these, you know, self-service and dashboards, I would say I personally back to the habit. I have really made a much more conscious decision. I don't need spreadsheets as much to make a point. I can actually present it in a completely different way because people engage with AI differently. And that's making a difference in how I'm approaching. And I like it myself when I tell a story and I feel like it resonates for the other person better. That doesn't quite need to be prepared. You don't need to prepare for multiple days. It's a lot more live. It's a lot more ready. It's a lot more self-service. That is a use case that I like the most.

Sabrina Ortiz: What's an example of something that you've been able to create using AI that's much more engaging or better at telling the story you want to tell than just looking at a spreadsheet?

Priya Srinivasan: It's a very internal use case, and I'm going to say this. It was actually around our cloud costs. And we have a tool called Apptio. And Apptio essentially is something that explains, hey, here's your cloud cost. It has full visibility, and it will tell you recommendations where you need to do it. So we have this data because we use, again, FinOps. We use all our technologies internally. So we use it and we actually have Apptio integrated with Turbo. Turbonomic is another tool of ours which can take recommendations. So we've done it. But any time I had to go and present to someone, I had to build a shadow of it. I have to build to say this is how our cloud costs looks like. These are the actions. Here's what we focused. I'm still building one chart. And I always felt like, how do I skip this step? Now that it took me a lot of time, I'm taking it from two tools and putting it as a single chart, I'm like, I want to be out of that, even building that chart. And this is where the Bob example from previously used Bob. And Bob helped generate always live chart view, always live, because it's ingesting data from two places. It's taking Apptio's information, the Apptio dashboard information and Turbonomic and everything else that we have to basically say, this is your single slide executive summary on it.

Sabrina Ortiz: Nice.

Priya Srinivasan: And so I don't need to build that single slide executive summary.

Sabrina Ortiz: Right. AI built it. And that wasn't the best use of your time anyways, right?

Priya Srinivasan: Correct. Mine was more around telling the story, the actions, and making sure I'm helping the team get towards that.

Sabrina Ortiz: The analysis.

Priya Srinivasan: The analysis. And I am spending more time on that actions rather than actually spending time on building the story.

Sabrina Ortiz: Well, thank you so much. I'm so happy we got to chat. And I think that's a perfect note to end on, because it's a perfect example of using AI in a way that makes your life just a bit easier and taking it in bite size steps. So, yeah, thank you so much for your time. Congrats on the launch. And yeah, I hope you have a great rest of your IBM think.

Priya Srinivasan: Thank you. Thank you for having me and appreciate your time as well.

Sabrina Ortiz: Thank you. Awesome. Thank you, guys.