AI x CIO: Benni Blau on the Changing Role of the CIO

Hello everyone, I'm Abhijit Mitra. I'm the CEO of Outreach and I'm very excited to be here today with Benni Blau Chief Process and Information Officer at SAP and we are going to talk about AI and the changing role of the CIO So welcome Benni to this podcast. Thanks for having me happy to be here so Yeah, well, I was thinking but you could start with With your title. It's Chief Process and Information Officer It's a very interesting title and I'm sure we would all love to know what that actually means Yeah, so maybe let me start from from that angle It's not because I like long titles, but the process is really it's really dear to my heart because I truly believe If you want to transform your business, it starts with processes and not with IT IT is a means to get to an outcome, but you need to start from a business lens basically and Especially in a software company and in high tech you might know it because everyone can do it There's a tendency to solve every problem with another tool or product And sometimes that's not needed and this is why we also the SAP did a major shift we pivoted basically from a product to a process centric IT which has major implications, but it really helps to basically align Ourselves to our business partners better so we can share business KPIs we can focus on outcomes not on classical IT KPIs And it really helps us to approach problems a little bit different That's awesome. So you're really thinking about how you apply technology to solve business problems And not the other way around which is exactly what I believe in as well Yeah, but that's the point right because it's it's a little bit of a chicken egg problem Typically you say process first IT second, but sometimes your business counterpart does not know what they don't know So you have to basically first get them in touch with technology, especially on the AI front To actually get them to know what's important what's in the space of their potential solution to their problem And that's that's what we also try to do. So we are not too formal about the sequence All right. So let's talk about AI and You and I were chatting about this just before we started this podcast and one of the things you told me is that you know, traditionally process is a very well structured step-by-step the one two three four kind of thing But you said that with AI a lot of that is changing. So how is that changing? Very like changing all of that And yeah, so if people talk about processes they typically think in you know Like predetermined sequences of tasks that someone has to perform in order to get after those sequences to a certain outcome, right? It's it's predetermined. You can well document it. You can audit it and it's all good with agentic, I think there will be a disruption to that kind of mindset because Typically what you want to do is you express an objective or some value you want to achieve and then something happens, right? There's an orchestrator agent. There are niche expert agents. They orchestrate themselves and Right from the beginning. You don't know what's what's going to happen. You might ask the same question With a different sequence happening to get to a different outcome. And I think that's the magic also We want to see because as an end user you have not you don't have to talk about the how you simply talk about the what And the rest happens based on uh on artificial intelligence and and the agentic layer so you're really talking about the vision where you have a business outcome that you want to get to And the AIi agents are helping you autonomously figuring out what the best next action should be for you to do So you can get to that outcome that is just amazing now One of the things that I hear from a lot of my customers I don't know if it's true for you as well is that For the ai agents to do what they have to do to take their decisions do that work um They need data And a lot of companies are struggling with getting good data together. So how How how are you handling that and what's your view on that? Yeah, so first of all for me it's important that we as IT are customer zero for everything SAP also provides to our customers and especially in the data space we recently launched business data cloud, um, and The idea is to solve a very common problem with all Large enterprise customers because they have a very Dispersed heterogeneous data landscape. They have SAP applications non-SAP applications a warehouse system and uh As a prerequisite for any good high value AI cases and especially agentic You need a semantically well-defined harmonized data layer with clear dominance um and with bdc you can basically Achieve that because it's uh, what we call select the heart of the concept is a data product So what we do is when we ship an application You get a data product out of the box, which is a semantically well-defined API that exposes a certain business data Object and it's the single point in your company where that is exposed. There's zero copy, right? It's uh, it can directly be used and then you have data sphere as the layer Where you compose those I would say raw Business data objects into more complex ones and this is then the foundation for Transactions for AI for agents and for analytical assets That's awesome. You know, we always learned Um in computer science that you know garbage in is garbage out so So and this I don't think any different with any technology that we have whether it's agentic generative or or whatever it is It's a some of the principles always holds true. I'm really excited to hear about that though um, so let's say now that you have somewhat figured out your data and then you have Autonomous agents who are actually working on their data figuring out the next best action So this process is actually this business outcome that you're trying to get to The thing that you talk about a lot i've heard you talk about this is focusing on quality over quantity and you said business ai is about quality or Talk a little bit about that. Tell us about that like What does it mean? Yeah, I think I mean when it all started I typically in many and I talked to my peers. I typically hear the question How many AI cases do you have? I think it's entirely the wrong question. It's because it It basically uh forces a certain volume of your cases and so on but you you lose track of A true value. So for us internally we have a very stringent value assessment process Where so I don't you know prevent any idea from coming up and being assessed don't get me wrong. Uh, but then For that idea to be processed and for us to really put investments in there's a very stringent value assessment and I would rather have like small amount of cases that yield true value for for us internally for the employees and the different functions for SAP then Having a broad spectrum of like I would say more of commodity cases. Yeah That where there is basically no return of investment So the ROI is really what you're looking for and where the ROI is maximum maximize that focus on a few get it right Then focus on the many and not have any focus at all. Really So when you talk about ROI though, um, is it always about like productivity or time spent or time saved or like what kind of ROI what kind of outcomes I know it depends on the use case, but what are the top things that you are thinking of in your mind? Yeah, I mean typically you have two categories. The one is as you said, right? productivity increase that yield or that that lower cost in certain area Uh need less manual steps and so on which is more of a bottom line Perspective, but then there's also clearly top line cases where there's more value for customers We have SAP for me as an example, right? So a single stop shop for customers when you see what you purchase support tickets everything you do via the single Central, uh, and then we put a lot of AI in for example search other capabilities Uh because this is a true value then for all our customers, right? So there's I guess top line Uh category and the bottom line category, uh in this value assessment both is important One of the things that I I look at when I use the I myself is that you know How much not just time it saves but like? Um how much of the grunt work that I don't have to do anymore Truly, I mean, um, yeah, so I mean we make software for go-to-market people, right? So Um, imagine the amount of research you have to do before Like say I video And how much of that can be done by AI? Um or how much of my proactive customer reach outs can actually be done by AI before I get involved So that's what like, you know for me personally that makes a huge difference. Um Um, and just so you know, Benny, today is the last day of our quarter and I'm here in a podcast with you That speaks speaks for your product and And the reason I can do that and I posted that on LinkedIn just before our broadcast is that Because my AI has already projected the quarter out for me and has been doing that consistently for the last couple of quarters Um, and it gets all this this information from all the interactions that's happening with our customers Um, so anyway that kind of saves us a lot of time saves me a lot of time to your point Right. I think that also answers the question of what's the the human in the formula and I mean this is for example, we do for our engineering folks a lot of AI capabilities automation and we take basically we take off all the nasty tasks that nobody wants to do like documentation, bug fixing, security scans, uh open source scanning cicd pipeline creation commit documentation everything we try to take off their plate So they can fully like they can transform how they work into more value-added Uh a task and I think it's the same for the sales space marketing And I think I think that transformation we see in the industry. So there is uh, I would rather talk about more outcome and transformation of higher value tasks Than anything else when there's a discussion between you know, what's the role of? People versus AI. Yeah, and and we hear a lot of that like there's a lot of concern, genuine concern around. Oh, what will we do if ai does everything for us? Now my point has always been that it's technology in the service of people And not the other way around. Um, and then the more technology we have the more productive we become and uh, that's sort of been true forever as far as I can remember like, you know, any new technology opens up new opportunities Um, do you believe the same as way? Uh, Benny or do you see it? Actually I like to think in top line and opportunities rather than input and bottom line. So From my point of view it just extrapolates the the potential of output Uh Rather than a discussion of do we need less or more people for something? It's just like you can do way more with the same Yeah And you can move people into more exciting roles and tasks in the company and I think that's the big The big upside. Um, yeah, by the way I mean as everyone is applying those technologies There's also a pressure for everyone to get more productive in terms of outcome, right? I think That's the big theme you will see and I think it's a I would rather look at it from the opportunity side than any doom scenario that some people put out there absolutely, um so one of the things that um We also see internally with an outreach and also with our customers that when you have Like what's happening out there in the market? Is there a lot of people saying that I have a I use my tool And so it gets a little confusing. Um for for you as a buyer as to okay Which one do I use for what purpose and stuff? And so I have seen some of our customers struggle with that Because they either end up buying too much or nothing because they're not sure like which one actually does anything at all So in that situation like when you have all the different technologies How do you decide or what's your criteria when you decide like how should you put it together? So you have sort of one conservative tech stack which makes sense for you for your yeah It's a multifaceted question. Um so I think um for example we as a ciO you're typically confronted with the problem of Multiple tools for different purposes People can like in the SAAS world like your employees can also purchase things they can try out things you have Like the land and expand motions everyone else does with us as well, right? Right? Yeah, so um So what we do on the enterprise architecture side is we use lean ix to basically model all the business capabilities We have to support with tooling And then we have an inventory of everything that runs in the company and then you immediately see where there's redundancies right, and we actively trigger the discussion of Consolidation of stack at some point in time, which doesn't mean we do not allow for trying out things You know, uh assessing the market. I mean that's that's like that's bread and butter for us as well But at some point in this in this life cycle You need to come to a point where things converge and we cannot afford You know too much things for the same supporting the same business capability especially on the platform side, um We rely heavily on btp and gen ai hub for for the cases we develop for and not because it's our product but because we want to make sure There's you know data protection is assured GPR is assured. Everything runs in a compliant and governed and Reliable fashion and and this is why I think the lower you are in the stack Especially on platform level you cannot afford a zoo of things if you go up up the stack in the Higher application space. I think it's okay to experiment and try out But at some point it has to converge also from cost point of view. Yeah, you mentioned governance You mentioned data consistency process consistency So we hear a lot of that from our top enterprise customers as well Um, and many of our CIOs are actually really struggling with a lot of these these issues So as a closing Question I would have for you is like, you know What advice would you have for other CIOs who are trying to leverage AI to you know Get to those outcomes that you are looking for that they're looking for But at the same time struggling with some of these problems about like, you know tech stack explosion You know a lot of the challenges we talked about today Quality of quantities and so forth importance of good data. So what what advice would you have for these CIOs to be more effective? I mean first and foremost and it's clear. I disrupts a lot of different classical roles in the company and so on but especially for the the CIO role it completely elevates the role to a completely different strategic level from my point of view so Of what maybe was a cost center back then in the days is now like if I look at SAP, right? so we we we have productivity promises to our customers of 30 percent now what does it mean in return is we have to live up to that promise also internally and My team our team is responsible for You know offering the tooling the capability enablement For everyone internally and the different functions in order to get to that level of productivity And that puts you in a I think in a Highly demanding but also in a very luxury position in the company and I would just recommend To everyone else live up to that To that expectation and you know have also a different Self image of the unit in the company internally and be part of the right strategic discussions rather than being a service provider Amazing that just sums it up so well And the title imagine the title of a podcast is AI and the changing role of the CIO And that's exactly how you summed it all up for all of us. So thank you so much Benny for being here with me today. I really appreciate. Um, I think our audience will appreciate this quite a bit when this goes out Um, thanks for being here with us and sharing that was fun. I happy to be back

Hello everyone, I'm Abhijit Mitra. I'm the CEO of Outreach and I'm very excited to be here today with Benni Blau Chief Process and Information Officer at SAP and we are going to talk about AI and the changing role of the CIO So welcome Benni to this podcast. Thanks for having me happy to be here so Yeah, well, I was thinking but you could start with With your title. It's Chief Process and Information Officer It's a very interesting title and I'm sure we would all love to know what that actually means Yeah, so maybe let me start from from that angle It's not because I like long titles, but the process is really it's really dear to my heart because I truly believe If you want to transform your business, it starts with processes and not with IT IT is a means to get to an outcome, but you need to start from a business lens basically and Especially in a software company and in high tech you might know it because everyone can do it There's a tendency to solve every problem with another tool or product And sometimes that's not needed and this is why we also the SAP did a major shift we pivoted basically from a product to a process centric IT which has major implications, but it really helps to basically align Ourselves to our business partners better so we can share business KPIs we can focus on outcomes not on classical IT KPIs And it really helps us to approach problems a little bit different That's awesome. So you're really thinking about how you apply technology to solve business problems And not the other way around which is exactly what I believe in as well Yeah, but that's the point right because it's it's a little bit of a chicken egg problem Typically you say process first IT second, but sometimes your business counterpart does not know what they don't know So you have to basically first get them in touch with technology, especially on the AI front To actually get them to know what's important what's in the space of their potential solution to their problem And that's that's what we also try to do. So we are not too formal about the sequence All right. So let's talk about AI and You and I were chatting about this just before we started this podcast and one of the things you told me is that you know, traditionally process is a very well structured step-by-step the one two three four kind of thing But you said that with AI a lot of that is changing. So how is that changing? Very like changing all of that And yeah, so if people talk about processes they typically think in you know Like predetermined sequences of tasks that someone has to perform in order to get after those sequences to a certain outcome, right? It's it's predetermined. You can well document it. You can audit it and it's all good with agentic, I think there will be a disruption to that kind of mindset because Typically what you want to do is you express an objective or some value you want to achieve and then something happens, right? There's an orchestrator agent. There are niche expert agents. They orchestrate themselves and Right from the beginning. You don't know what's what's going to happen. You might ask the same question With a different sequence happening to get to a different outcome. And I think that's the magic also We want to see because as an end user you have not you don't have to talk about the how you simply talk about the what And the rest happens based on uh on artificial intelligence and and the agentic layer so you're really talking about the vision where you have a business outcome that you want to get to And the AIi agents are helping you autonomously figuring out what the best next action should be for you to do So you can get to that outcome that is just amazing now One of the things that I hear from a lot of my customers I don't know if it's true for you as well is that For the ai agents to do what they have to do to take their decisions do that work um They need data And a lot of companies are struggling with getting good data together. So how How how are you handling that and what's your view on that? Yeah, so first of all for me it's important that we as IT are customer zero for everything SAP also provides to our customers and especially in the data space we recently launched business data cloud, um, and The idea is to solve a very common problem with all Large enterprise customers because they have a very Dispersed heterogeneous data landscape. They have SAP applications non-SAP applications a warehouse system and uh As a prerequisite for any good high value AI cases and especially agentic You need a semantically well-defined harmonized data layer with clear dominance um and with bdc you can basically Achieve that because it's uh, what we call select the heart of the concept is a data product So what we do is when we ship an application You get a data product out of the box, which is a semantically well-defined API that exposes a certain business data Object and it's the single point in your company where that is exposed. There's zero copy, right? It's uh, it can directly be used and then you have data sphere as the layer Where you compose those I would say raw Business data objects into more complex ones and this is then the foundation for Transactions for AI for agents and for analytical assets That's awesome. You know, we always learned Um in computer science that you know garbage in is garbage out so So and this I don't think any different with any technology that we have whether it's agentic generative or or whatever it is It's a some of the principles always holds true. I'm really excited to hear about that though um, so let's say now that you have somewhat figured out your data and then you have Autonomous agents who are actually working on their data figuring out the next best action So this process is actually this business outcome that you're trying to get to The thing that you talk about a lot i've heard you talk about this is focusing on quality over quantity and you said business ai is about quality or Talk a little bit about that. Tell us about that like What does it mean? Yeah, I think I mean when it all started I typically in many and I talked to my peers. I typically hear the question How many AI cases do you have? I think it's entirely the wrong question. It's because it It basically uh forces a certain volume of your cases and so on but you you lose track of A true value. So for us internally we have a very stringent value assessment process Where so I don't you know prevent any idea from coming up and being assessed don't get me wrong. Uh, but then For that idea to be processed and for us to really put investments in there's a very stringent value assessment and I would rather have like small amount of cases that yield true value for for us internally for the employees and the different functions for SAP then Having a broad spectrum of like I would say more of commodity cases. Yeah That where there is basically no return of investment So the ROI is really what you're looking for and where the ROI is maximum maximize that focus on a few get it right Then focus on the many and not have any focus at all. Really So when you talk about ROI though, um, is it always about like productivity or time spent or time saved or like what kind of ROI what kind of outcomes I know it depends on the use case, but what are the top things that you are thinking of in your mind? Yeah, I mean typically you have two categories. The one is as you said, right? productivity increase that yield or that that lower cost in certain area Uh need less manual steps and so on which is more of a bottom line Perspective, but then there's also clearly top line cases where there's more value for customers We have SAP for me as an example, right? So a single stop shop for customers when you see what you purchase support tickets everything you do via the single Central, uh, and then we put a lot of AI in for example search other capabilities Uh because this is a true value then for all our customers, right? So there's I guess top line Uh category and the bottom line category, uh in this value assessment both is important One of the things that I I look at when I use the I myself is that you know How much not just time it saves but like? Um how much of the grunt work that I don't have to do anymore Truly, I mean, um, yeah, so I mean we make software for go-to-market people, right? So Um, imagine the amount of research you have to do before Like say I video And how much of that can be done by AI? Um or how much of my proactive customer reach outs can actually be done by AI before I get involved So that's what like, you know for me personally that makes a huge difference. Um Um, and just so you know, Benny, today is the last day of our quarter and I'm here in a podcast with you That speaks speaks for your product and And the reason I can do that and I posted that on LinkedIn just before our broadcast is that Because my AI has already projected the quarter out for me and has been doing that consistently for the last couple of quarters Um, and it gets all this this information from all the interactions that's happening with our customers Um, so anyway that kind of saves us a lot of time saves me a lot of time to your point Right. I think that also answers the question of what's the the human in the formula and I mean this is for example, we do for our engineering folks a lot of AI capabilities automation and we take basically we take off all the nasty tasks that nobody wants to do like documentation, bug fixing, security scans, uh open source scanning cicd pipeline creation commit documentation everything we try to take off their plate So they can fully like they can transform how they work into more value-added Uh a task and I think it's the same for the sales space marketing And I think I think that transformation we see in the industry. So there is uh, I would rather talk about more outcome and transformation of higher value tasks Than anything else when there's a discussion between you know, what's the role of? People versus AI. Yeah, and and we hear a lot of that like there's a lot of concern, genuine concern around. Oh, what will we do if ai does everything for us? Now my point has always been that it's technology in the service of people And not the other way around. Um, and then the more technology we have the more productive we become and uh, that's sort of been true forever as far as I can remember like, you know, any new technology opens up new opportunities Um, do you believe the same as way? Uh, Benny or do you see it? Actually I like to think in top line and opportunities rather than input and bottom line. So From my point of view it just extrapolates the the potential of output Uh Rather than a discussion of do we need less or more people for something? It's just like you can do way more with the same Yeah And you can move people into more exciting roles and tasks in the company and I think that's the big The big upside. Um, yeah, by the way I mean as everyone is applying those technologies There's also a pressure for everyone to get more productive in terms of outcome, right? I think That's the big theme you will see and I think it's a I would rather look at it from the opportunity side than any doom scenario that some people put out there absolutely, um so one of the things that um We also see internally with an outreach and also with our customers that when you have Like what's happening out there in the market? Is there a lot of people saying that I have a I use my tool And so it gets a little confusing. Um for for you as a buyer as to okay Which one do I use for what purpose and stuff? And so I have seen some of our customers struggle with that Because they either end up buying too much or nothing because they're not sure like which one actually does anything at all So in that situation like when you have all the different technologies How do you decide or what's your criteria when you decide like how should you put it together? So you have sort of one conservative tech stack which makes sense for you for your yeah It's a multifaceted question. Um so I think um for example we as a ciO you're typically confronted with the problem of Multiple tools for different purposes People can like in the SAAS world like your employees can also purchase things they can try out things you have Like the land and expand motions everyone else does with us as well, right? Right? Yeah, so um So what we do on the enterprise architecture side is we use lean ix to basically model all the business capabilities We have to support with tooling And then we have an inventory of everything that runs in the company and then you immediately see where there's redundancies right, and we actively trigger the discussion of Consolidation of stack at some point in time, which doesn't mean we do not allow for trying out things You know, uh assessing the market. I mean that's that's like that's bread and butter for us as well But at some point in this in this life cycle You need to come to a point where things converge and we cannot afford You know too much things for the same supporting the same business capability especially on the platform side, um We rely heavily on btp and gen ai hub for for the cases we develop for and not because it's our product but because we want to make sure There's you know data protection is assured GPR is assured. Everything runs in a compliant and governed and Reliable fashion and and this is why I think the lower you are in the stack Especially on platform level you cannot afford a zoo of things if you go up up the stack in the Higher application space. I think it's okay to experiment and try out But at some point it has to converge also from cost point of view. Yeah, you mentioned governance You mentioned data consistency process consistency So we hear a lot of that from our top enterprise customers as well Um, and many of our CIOs are actually really struggling with a lot of these these issues So as a closing Question I would have for you is like, you know What advice would you have for other CIOs who are trying to leverage AI to you know Get to those outcomes that you are looking for that they're looking for But at the same time struggling with some of these problems about like, you know tech stack explosion You know a lot of the challenges we talked about today Quality of quantities and so forth importance of good data. So what what advice would you have for these CIOs to be more effective? I mean first and foremost and it's clear. I disrupts a lot of different classical roles in the company and so on but especially for the the CIO role it completely elevates the role to a completely different strategic level from my point of view so Of what maybe was a cost center back then in the days is now like if I look at SAP, right? so we we we have productivity promises to our customers of 30 percent now what does it mean in return is we have to live up to that promise also internally and My team our team is responsible for You know offering the tooling the capability enablement For everyone internally and the different functions in order to get to that level of productivity And that puts you in a I think in a Highly demanding but also in a very luxury position in the company and I would just recommend To everyone else live up to that To that expectation and you know have also a different Self image of the unit in the company internally and be part of the right strategic discussions rather than being a service provider Amazing that just sums it up so well And the title imagine the title of a podcast is AI and the changing role of the CIO And that's exactly how you summed it all up for all of us. So thank you so much Benny for being here with me today. I really appreciate. Um, I think our audience will appreciate this quite a bit when this goes out Um, thanks for being here with us and sharing that was fun. I happy to be back