2927: SAP's Data Sphere: Redefining AI and Data Management

[00:00:00] What does the future hold for data and analytics, especially in this age dominated by artificial

[00:00:08] intelligence? Well, today I'm delighted to welcome Irfan Khan, Chief Product Officer

[00:00:13] at SAP onto the podcast. And today's episode, we'll explore how SAP Datasphere is introducing

[00:00:21] the concept of a business virtual data fabric, allowing the data federation without the need

[00:00:28] to physically move data. And Irfan will shed light on the critical role of preserving data

[00:00:34] context and metadata for training high quality generative AI models. I also want to talk

[00:00:41] a little about how SAP is leveraging knowledge graphs and vector stores for advanced AI capabilities

[00:00:48] and talk about, of course, the importance of data quality, lineage and ethical usage

[00:00:54] to ensure a responsible deployment of generative AI, along with a few exciting real world use

[00:01:00] cases and stories along the way. So buckle up and hold on tight as I beam your ears all

[00:01:07] the way to the UK so you can join me and Irfan as we go on a journey into the cutting edge

[00:01:12] world of data and AI with some pretty inspiring insights from a true industry leader. But

[00:01:18] enough from me. Let's get Irfan on the show now. So a massive warm welcome to the show.

[00:01:25] Can you tell everyone listening a little about who you are and what you do?

[00:01:29] Well, thank you so much. Firstly, I appreciate the opportunity to connect with you today.

[00:01:33] If it was a radio station, I would be first time caller but long time listener. I have

[00:01:37] listened to your podcast for some time and I know you have some very broad spectrum

[00:01:42] of discussion. So a little bit about myself, Irfan Khan. I'm at SAP now for about 12 plus

[00:01:48] years, arrived through an acquisition, a company called Sybase where I was there for 18 years.

[00:01:54] So in my entire career, I've effectively spent most of that time in data and analytics. My

[00:01:59] current role and responsibility is that I'm the Chief Product Officer of SAP's Data Analytics

[00:02:04] Technology. So I'm running a development organization. Very fortunate to work with some very talented

[00:02:10] colleagues and we're working on the cutting edge. So all the new innovations around AI

[00:02:15] and force planning and predictive and all the other pieces that you would associate with

[00:02:19] data analytics, we've been working on for quite a number of years now.

[00:02:23] And as a long time listener, first time caller, you know one of the things I'm fascinated

[00:02:27] by on this podcast is my guest's origin story. So I'd love to find out a little bit more

[00:02:32] about how you got into your current role. Can you share how you got into the tech industry,

[00:02:37] where that passion for tech came from? Maybe something lit the spark, but there's always

[00:02:41] a story there, right?

[00:02:42] Oh yeah, for sure. I mean, I think my ambition in technology goes way back and I've got to

[00:02:48] maybe attribute that to my father who sadly passed away about a year plus ago now. But

[00:02:53] he certainly was a very significant, you know, not just a role model, but also one who gave

[00:02:58] a lot of guidance. And, you know, I remember going back in the early, early days, you know,

[00:03:02] back into school, starting off with the first batch of, I guess I was in the first batch

[00:03:06] of students that were given the opportunity to formally get into computer studies and

[00:03:10] computer science. Right? So the ambition was certainly there from the beginning. I didn't

[00:03:14] know what career could be, could be had in this domain 30 plus years ago. But as I look

[00:03:19] back now, I mean, the origin story I think will be driven by two key, key factors. The

[00:03:23] first one being having a keen interest and really being motivated by technology. And

[00:03:28] the second one was really finding myself landing on my feet with one significant role at Sybase

[00:03:34] where I was the chief technology officer and then landing in SAP, an applications company.

[00:03:38] So completely different departure from what I was used to from pure technology into the

[00:03:42] application domain. I was fortunate enough to really have another rebirth as a career.

[00:03:46] I went into sales, right? Which sounds a little odd from having spent all my life in, you know,

[00:03:51] technology and product. But I spent seven of the, or actually probably eight of the last

[00:03:55] 12 plus years in sales, right? So that in itself gives you a very substantial point of view

[00:04:00] around what customers really care about, what their motivation is and ultimately what does it

[00:04:05] take to take a product into interconception by new customers. So a lot of that is really

[00:04:09] grounding me in terms of some of the challenges and opportunities that we see on the horizon in

[00:04:14] SAP. And I don't want to date you here, but if I take you back one more time to those school days,

[00:04:19] what was the computers that you were using in the school? For me, I'm a very old guy now. So

[00:04:24] it would have been a BBC micro with 32K. Yeah, so if you remember, there was a kind of a big sort

[00:04:31] of standoff between Amstrad and BBC micros, right? So I was in the Amstrad camp, CPC 464 to be

[00:04:38] precise, right? So it was a very ancient thing. It had a tape cassette recorder in it. So, you

[00:04:44] know, you'd go and copy the games with your friends. I shouldn't say that, right? But you

[00:04:46] copied the games and you stick them into the tape recorder and only find out at the 98% of loading

[00:04:52] the game that there was some error, right? So I feel the real first world problems that you and I

[00:04:56] went through.

[00:04:57] 100%, those tape dropouts, man. I could have an old podcast episode dedicated to that. But of

[00:05:04] course, it would be that path, although the computing back then was quite primitive to what we

[00:05:09] have now, that path would eventually lead you to SAP. So can you tell me more about the business

[00:05:14] and the problems that you're addressing for your customers at the moment?

[00:05:18] For sure. I mean, maybe just building on what you said. I mean, the historical compute

[00:05:23] foundations, the compute and store foundations, I remember my first, I would call it proper

[00:05:27] computer that I ended up doing my final year project on as I was an undergraduate at the time.

[00:05:31] It had four meg of main memory and 20 gigs of storage. And that was able to run at the time. It

[00:05:38] was a project that was running on Oracle actually at the time, funnily enough. And we were able to

[00:05:42] run the entire RDB mess in four meg, right? With a 20 megabytes worth of storage. So look, I think

[00:05:47] if you look at SAP today, we're at the pioneering end of business applications. I mean, the

[00:05:51] company celebrated its 50th anniversary not too long ago. And in these five decades, it's really

[00:05:56] supplanted itself as really being the world leader in business applications. And you can pick any

[00:06:01] industry, whether that's going to be healthcare, or it could be financial services, or

[00:06:06] pharmaceutical, SAP has a real stronghold in 20 plus industries. And if you imagine companies run

[00:06:12] their businesses in various business applications, SAP maps its business applications to the entire

[00:06:18] company. So front office, middle office, back office. And you could imagine that the level of

[00:06:22] robustness that you need to have if you're a Centrica, or if you're a Nestle, or if you're a

[00:06:28] McLaren, or whoever it might be, these are very large organizations and very different business

[00:06:33] models. And SAP maps its business applications to marry up to the exact requirements, the demands

[00:06:39] of those businesses. So very, very interesting sort of to look at the applications evolution of

[00:06:43] SAP. But I think right now, the problem that we're solving is the evolution, and we'll get into some

[00:06:48] of the GNI topics in just a moment. But those first class business processes that have been

[00:06:52] robust and running for decades now have to evolve. And then when they evolve, of course, there's two

[00:06:58] levels of efficiencies that customers want, they want the price to come down in terms of running

[00:07:02] those operations and the applications. But equally, they want the value of those applications to

[00:07:06] really guide them, almost like the front headlights of a car, making sure that they point in the

[00:07:11] right direction and help them really navigate the twisty roads ahead as well.

[00:07:16] It's amazing just how much has changed. I mean, you mentioned four meg of RAM there. And for a lot

[00:07:20] of people, that's like a photo on their phone. And I'm looking at the iPad Pro recently, and the

[00:07:25] latest version has got 16 gig of RAM, which is more than a MacBook. It's just phenomenal how

[00:07:31] quickly it's increasing, isn't it? And how we take that compute power for granted all that.

[00:07:36] Absolutely. I mean, if you imagine that the top level iPhone now has a terabyte worth of storage,

[00:07:42] which isn't physical sort of spinning disks anymore, it's all in memory storage effectively. So it's

[00:07:47] incredible to think how much efficiencies have been gained and the level of micro architectures that

[00:07:53] we now are in the palm of our hands.

[00:07:55] 100%. And if we fast forward to present day, of course, we're now specifically here today to talk

[00:08:01] about SAP's data and analytics portfolio. And of course, the latest edition of SAP's DataSphere. So

[00:08:08] for anyone that's not familiar with that, can you just give me a rundown of exactly what it is and

[00:08:13] ultimately what you're aiming to achieve? What kind of problems you're solving here?

[00:08:17] Yeah, so Neil, if you look back over the last several decades, we've seen a substantial maturing

[00:08:22] in the data and the analytics space. But there is still a lot of old school mentality and practices

[00:08:29] that go on. And as an example, I'll give you a practical home example. And if you were today about to move

[00:08:35] home, you probably go through a process where you have to go and look for all of the different sort of, I

[00:08:39] don't know, all that paperwork that you may have, whether that's in your filing cabinet, it could be maybe

[00:08:44] stuffed in the kitchen drawer, it could be bills that you paid, and you still want to have some records for

[00:08:48] and that could be up in your attic. And you tend to have to look at this data in some historical context,

[00:08:53] but more often than not, businesses have to have that access, whether it's for regulation or

[00:08:58] compliance purposes, but they don't always know the reliability of the data. And sometimes they don't

[00:09:02] know where the data physically is. And if we look at the evolution of data architectures, some advocates

[00:09:08] and meaning some vendors have advocated for building a very large filing cabinet, let's call it a data lake

[00:09:13] today. And they're asking people to move all of their individual silo different sort of data sets and

[00:09:19] physically move them into this new large filing cabinet. That comes with a level of frustration, cost,

[00:09:26] and at the same time, a lot of heavy lifts and not a lot of people have the aptitude or the desire to want to

[00:09:31] become data janitors, right where you have to go look for all the different pieces of information. And if you

[00:09:36] marry that sort of practical home example now into the enterprise, where you do find large organizations

[00:09:41] constantly looking at trying to drive efficiencies and storage and rationalizing all the different data

[00:09:47] sets, what Datasphere introduces into the market is a concept of a business virtual data fabric. And this is a

[00:09:53] fabric will describe itself. I mean, the physical fabrics are typically stitched together with multiple

[00:09:59] different pieces, but you don't always have the ability to move physically all the different attributes, all

[00:10:03] the different data sets into one location. So it's built around the concept of federation. And what is

[00:10:09] federation? Well, it means that you don't physically need to move the data, you can access it remotely. And

[00:10:14] there's a couple of trends, I think, that are supporting this now where if you look at the manner in which

[00:10:18] infrastructure is evolved, you look at the hyperscalers, or which, you know, you consider Microsoft, of course,

[00:10:24] Google and Amazon, right as very significant players in the in the space of infrastructure, they have introduced

[00:10:31] very substantial levels of compute and store and infrastructure around networking. And those networks are

[00:10:37] becoming very significant and low latency based as well. So it's actually a lot more accessible now to look at

[00:10:42] data in multitude of different locations, you don't need to have all data physically available in one location.

[00:10:48] So the notion of federation is very key. And the second part is that when you start building next generation of

[00:10:53] applications, you want to build the foundations of the technology once and build the applications on top of that not

[00:11:00] constantly having to evolve the next filing cabinet and the next version of the filing cabinet as it may be. And this is

[00:11:05] exactly where we see data sphere really resonating well with the market. We have over 1400 customers already, and

[00:11:11] they're building the next generation data foundations on data sphere, networking within the open ecosystem, which I'll

[00:11:17] come on to in just a moment. But that kind of hopefully gives you a bit of a summary around why we're excited about

[00:11:21] data sphere and how it maps to the existing IT landscape.

[00:11:25] And one of the things that put you guys on my radar was a large part of SAP's data sphere proposition is centered around

[00:11:32] the power of generative AI. And I think we must have broke a record on a tech podcast for going 15 minutes without

[00:11:39] talking about gen AI. But can you tell me more about that, and also the importance of utilizing contextual data?

[00:11:46] Absolutely. So let's start with contextual data, because that's a very good starting point. I mean, the majority of data

[00:11:52] that you have, because of this, the extract and moving data from one environment to another environment typically will

[00:11:58] result in you losing that metadata or the context. Take for example, in a business application like SAP, where you may

[00:12:05] have financial data or ledger data, the concept of an invoice, for example, and the minute that you move that data from

[00:12:12] the source system, and it doesn't need to be SAP, it could be any source system, and you copy that data, you typically

[00:12:17] lose those contextual parts. With data sphere, maintaining context and ensuring that the metadata is preserved really

[00:12:25] helps itself towards model training. So in the AI space, in particular, quality of data, and having well informed

[00:12:32] models and building models that are not going to hallucinate left, right and center is a very critical part. And with the

[00:12:38] benefits now with the data sphere, we've really put the two additional components within that the concept of a knowledge

[00:12:44] graph, which is a very significant component that most gen AI foundations were built around. And also vectorization,

[00:12:52] vector stores, right, which are there to support tokenizing and making sure that you can vector into the context of the

[00:12:57] data. So imagine that you've got looking for similarity of data within a very large data set, both the combination of the

[00:13:03] vector store and also the the knowledge graph will give you the means to be able to have much more clearer, coherent,

[00:13:10] generative AI models created or access. So this is another key element of why data sphere is really building itself as a

[00:13:17] credible capability, right and foundation for the next generation of it.

[00:13:21] And I think there is often more hype around the technology itself and some of the problems that businesses are aiming to solve

[00:13:29] here. So just to bring to life what we're talking about, are there any examples you can share of customers that have already

[00:13:34] embraced this way of thinking and seeing real business value seeing measurable results?

[00:13:40] Do if you take a look at the supply chain, majority of supply chains today are built around some signal management, right? So

[00:13:46] you have to look for disruptions, let's assume the Suez Canal as it was recently reported, you end up with challenges, right

[00:13:52] where now you've got this huge backlog convoy effect and your supply chains have been disrupted. The fact is that before

[00:13:59] that information was actually even publicly available, meaning to the to the news agencies, lots of disruptions would have

[00:14:05] been noticed already within the supply chain, there would have been probably some knock on effects, maybe manufacturing

[00:14:10] plants are already finding that there was a scarcity of different components. So in the gen AI world, right, you'd want to be

[00:14:16] able to link all those signals together and build models around that and then train those models. So as and when you see

[00:14:21] inference changes in sort of market climates and conditions, you can actually start behaving differently. And so therefore, I

[00:14:28] would look at the entire supply chain, taking a look at the whole foundation in processing around manufacturing is one

[00:14:33] example in industry, there is huge levels of advancement that have been made now certain customers, I can't name directly,

[00:14:39] but they've already started looking at how is it that they should start informing their manufacturing processes around making

[00:14:45] sure that they can interpret those signals. But from not from a traditional sense where you have to run a report at the end of

[00:14:51] the day and find out that there's a disruption that you may have a week from now, you want to be able to look at this almost

[00:14:56] in near real time. So there is a whole whole level of new dimension of use cases that can be built if you have the right

[00:15:02] level of data with the right level of granularity and making sure that you can have the hopefully the lack of

[00:15:08] hallucination that goes on so you can really rely and trust in the decisions that you can make around this infrastructure

[00:15:13] now.

[00:15:15] And I think decision making with gen AI can only be as strong as the data that feeds the technology. And I suspect that everybody

[00:15:22] listening has encountered the garbage in garbage out problem at some point in time. So what are some of the risks businesses

[00:15:28] face if that data isn't up to scratch? I suspect this is a question you get a lot.

[00:15:34] Yeah, I mean, Neil, if you take a look at the majority of it, it's always got a handbrake. And by handbrake, I mean, of a kind of

[00:15:39] virtual handbrake and that you can people can visualize. And that handbrake is typically a day's worth of old data before you

[00:15:45] make a decision. Making decisions against real time data, certainly this reason, I will probably point out maybe financial

[00:15:51] services where they have evolved, they probably got less of a forced handbrake, because of the level of investment or

[00:15:58] capabilities that they baked in to some of the decision making. And whereas majority of industries today, they look at

[00:16:03] historical data, and they try to make sure that they can forecast and predict for the future. The foundation of quality and

[00:16:10] data is really an industry in its own segment, you take a look at data, you look at master data, or you take a look at data

[00:16:16] quality, these attributes are typically there and have been around for a long time. And now in the in the space of generative AI,

[00:16:22] it sort of almost puts a turbocharger on that. Because as you said, garbage in garbage out, you want to minimize the amount of

[00:16:28] data, sort of quality issues that you have duplication or data, duplicating data, as the case may be, ensuring that you have a

[00:16:36] very clear lineage of the data knowing where it came from, what's the data of a reliable quality source. And typically, what

[00:16:43] will happen is that if you end up with the father, the way that you get from the source of the data, meaning that you took an

[00:16:48] extract, and that extract was put into another system, which was duplicated into another system, you know, it's almost like a

[00:16:53] family tree, the father out that you get from the family tree, you can't really have a strong level of understanding of the of the

[00:16:59] parents and the and the historical, the hierarchical view of where the data came from. So there is a lot that's going on. And we are

[00:17:05] certainly developing lots of fail safes, we have our ethical models around gen AI, we're putting a lot of emphasis in making sure

[00:17:12] that the quality of the data is never compromised, ethically. And of course, from a cultural cultural point of view as well,

[00:17:18] making sure there's no cultural biases in that data, a lot of stuff is going on in order to make sure that we can really build upon the

[00:17:25] foundation of this new, you know, mega trend that's upon us.

[00:17:29] And it's such an important topic, because I think last year, there was a lot of concern around sensitive corporate data being used to

[00:17:35] train LLMs. And for the most part, we've overcome a lot of those problems. But those concerns are still very real for a lot of business

[00:17:43] leaders listening. So how do you manage the data privacy risks and generative AI? And what are some of the safeguards that need to be in

[00:17:49] place to be able to use this technology responsibly? And as you said, ethically,

[00:17:54] I recently listened to one of your podcasts, in fact, it was around about the financial transformation and how gen AI can be used to be

[00:17:59] able to support that. And that's probably a good starting point. Because if you take a look at the the financial data, and looking at it

[00:18:06] more or less from a point of view of integrity and coherency, and making sure that data is, is not going to be misleading to shareholders or

[00:18:14] even internal C suite executives. And you now multiply that out into into society, right? Where the vast majority of me looking at this

[00:18:22] case right now, which is a very sad situation in the NHS, the National Health Service in the UK, where you know, the whole of the blood

[00:18:28] tests and all the other stuff that got contaminated. Maybe that's a historical maybe reminder, but it's one that's still ever present today.

[00:18:35] The data privacy aspects, whether it's patient client information coming from a variety of different sources, it's putting an even higher level

[00:18:43] of integrity around that for most of our corporations, because nobody can can actually bear having one of those public scandals, okay, right

[00:18:51] on their balance sheet, it just will they will not be able to survive. And whereas before people were sort of ride them through and people

[00:18:57] will be probably less, maybe tolerant of maybe or not tolerant. So they'd be more more acceptant to some of those changes. The reality is

[00:19:04] right now that you don't get a second chance. So whereas before you could sort of ride it through and have probably less checks and

[00:19:10] balances, now, everybody will have to have a high degree of conformity. And if I look through the lens of SAP, we actually have put a lot of

[00:19:16] robustness and rigor behind these processes. You know, we have before we generate any AI for any of our packaged applications, embedded AI has

[00:19:25] to run through our ethical committee to make sure that we're not violating any of the context around the quality or the, you know, the

[00:19:32] integrity of that data, there's a lot that goes on before a data can actually all the data models can actually be published. So I would assume

[00:19:38] that most companies of repute, repute and of course, value will be employing employing similar things on the horizon as well.

[00:19:47] And there's so many exciting things happening at the moment. There was the open AI video last week, there's talk of it being integrated into

[00:19:53] Siri possibly in the future. And you also put it right at the heart of the eye of the storm almost here. So in terms of data management, what

[00:20:01] excites you about the future? And are there any other insights that you can leave us with today or anything that excites you in particular?

[00:20:08] You're right. I mean, if you take a look at the open AI sort of announcements around chat UBT 4.0, and then you look more importantly at what

[00:20:15] Google is then doing in response and what Meta is doing in response. And this is where I think it becomes almost going back to the early 90s and

[00:20:23] the mid 2000s, where I was very much deeply into the data space. Knowing exactly like for example, if Sybase was to make an announcement, you could

[00:20:31] almost guarantee that the next next tech event, Microsoft and Oracle would follow suit. And I think innovation was really thriving at the time.

[00:20:38] Arguably over the last probably five to seven years, I wouldn't say things have been stagnant because you'll need to look at some of the big tech

[00:20:44] startups that have really flourished. But if I look at what's exciting me right now is what's next? I mean, you look at Google all of a sudden now

[00:20:51] they're really trying to get back into the into the mindset of people, not just from a pure search perspective, where they've had a dominance for decades now.

[00:20:59] Take a look at some of the things that they announced. They announced something called Gemini Live, for example, which is chatbot where you have a serious

[00:21:07] like assistant or whether it's going to be Alexa, but they tend to be one way traffic coming from those assistants, you can't really stop them, pause them,

[00:21:13] interrupt them and say, Hey, look, you know, I need more context. I don't understand what you're saying here. And having that two way dialogue with Gemini

[00:21:19] Live as it was announced from Google is very interesting. And then they started using things like, you know, for the creators, I think there was

[00:21:26] something called Google VO, they announced, right where it can generate a 1080p video for you at a high level of quality, you can influence it, what do

[00:21:34] you want the backdrop to have and generate a very substantial, you know, video sort of creative thing within a minute, right? A minute worth of

[00:21:41] content, lots of things are on the horizon. But that's all kind of in the in the space where individuals and your general purpose users will take

[00:21:48] advantages. What's exciting to me is really what SAP is planning and bringing it home, I guess, if I may, you know, maybe a bit of a plug here, so to

[00:21:56] speak on what's on the horizon for SAP. With business applications today, one of the fundamentals is that you rely upon business applications to make

[00:22:04] key decisions. But those decisions typically, as I've described, have become very much part of the fabric of your of your of the way that you run your

[00:22:11] business, the way you want to grow and expand your businesses, this is where Gen AI will really help you a lot, right in terms of being able to

[00:22:18] establish some very clear indicators of success. And for example, if you take a look at recruitment, and the way that recruitment works today, more

[00:22:26] often than not, you go through a very, you know, detailed process, you end up with candidates, you bring them in. And sometimes, you know, the the

[00:22:32] whole impression of first impressions or lasting impressions, that doesn't hold true in them. In this timeframe, right, you want to make sure that the

[00:22:38] majority of people that you bring in are going to be able to help you drive innovation to the next level. And what does that mean? Well, it's a certain

[00:22:45] profile of person that you need. So in our success factors business, for example, introducing generative AI, and having a means of being able to really

[00:22:53] help screen candidates ethically, and from a point of view, making sure that the data that's being used is, is then informing the foundations of

[00:23:00] success for the future, looking at it from a business applications perspective, having access to all of the data that may be sitting within your within

[00:23:07] your corporation, and building models and training models and having context, there's so much that we can now do with business applications that wasn't

[00:23:14] achievable before. And when you add in partnerships, like you know, whether it's going to be Nvidia that give you even greater levels of performance

[00:23:20] around the same data and training data, there's a lot that's going on. So I think Neil, I would summarize and say that lots going on in the industry

[00:23:27] right now, lots of moves by the large players, lots going on in open source, which I think like whether it's meta with Lama three, their contributions

[00:23:34] that they're making, there's going to be a lot on the menu of CIOs and CDOs in the future. And that's probably another podcast in its own on making. So as I

[00:23:42] said, first time caller, but longtime listener, and I'm looking forward to hearing more from you in the future on these topics.

[00:23:48] And it's there is an open invitation for you to come back and discuss some of these topics as well. And just listening to you that I mean, keeping up

[00:23:56] the pace with the speed of technology can be incredibly overwhelming for a lot of people listening. And the last few weeks, although we've heard

[00:24:02] about text to video, open AI is announcements, phones and tablets that are more powerful than laptops is a long, long way from those BBC micros

[00:24:12] 32k of RAM and there is this pressure to continuously learn now. So I've got to ask, where or how do you self educate? How do you keep up to

[00:24:20] speed with the trains?

[00:24:22] Yeah, I think once upon a time we would go on these structured courses, it was a five day course on topic A, and then you'd come back and then you

[00:24:28] earned the sort of the foundational principles and go on to course B, which would be another three days, that that whole historical learning

[00:24:36] was probably ready for the time or acceptable for the time, but now it's all just in time. YouTube videos are a very good source. I mean, great

[00:24:43] content providers like yourself will create sort of snippets of content that you can use and then build a talk track around and build a

[00:24:51] foundation of knowledge around. So my self learning journey is driven by two components. One is lots of interactions and networking within

[00:24:58] within corporations, like you know, of course, the customers that we work with, but equally so internally with some very, very bright minds that

[00:25:04] we're fortunate enough to have at SAP where, you know, a lot of brainstorming goes on. But it's a bite sized learning digestive right now.

[00:25:11] I mean, like if I look at the readers digest of the past, it was, you know, something that was end to end and you know, cover to cover that

[00:25:17] you had to read. Now, you get your information in very small bursts of packages, we're living in the course, the TikTok generation, right? And

[00:25:23] content and training and learnings comes in those same size packages. So lots of lots sort of uncoordinated searching and networking where you

[00:25:31] start from point A, it maybe it's a YouTube video set that you're looking at. And you could end up probably looking in the Amazon jungle right

[00:25:37] before you even notice that you were there. But this is all sort of leads towards the, you know, the the unsorted choreographed training and

[00:25:43] learning I think which for me works really, really well.

[00:25:46] And we've covered so much about your life in this 30 minute podcast interview today, from your school years, how your father inspired you

[00:25:55] and the computers you were using at school, etc, to where we are now. And of course, as we now come full circle, I'm gonna ask you to look back

[00:26:02] at your entire career because none of us are able to achieve any degree of success without a little help along the way. And very often we

[00:26:09] encounter people that see something in us invest a little time in us. So is there a particular person that you're grateful towards who maybe

[00:26:16] helps you get where you where you are that we could give a little shout out at the end of our conversation today?

[00:26:21] And they'll thank you for the opportunity to maybe highlight I mean, I if I was to put down who has helped me along my career would be a long

[00:26:28] list. And obviously, there will have to be a top five or a top 10, you got to make the cut at some point in time. And I would say without

[00:26:33] hesitation, I mean, one particular person who steps who stands out for me is Dr. Raj Nathan, he was a both a mentor and also a very significant

[00:26:42] leader that that I was able to use to to shape my own leadership credentials. And Dr. Raj Nathan and I worked together for almost a decade at

[00:26:50] Sybase in my formative years really working, understanding exactly the domain of technology, what's possible. And then he had a lot of

[00:26:57] leadership qualities, you know, I remember just I'll give you one incident just to sort of wrap this up on maybe on a more of a personal

[00:27:02] note. He was at a very lofty height, he was part of the executive leadership team of Sybase at the time, I'm a lowly sort of individual, he and I

[00:27:09] were traveling together. He had, of course, you know, first class or business class ticket at the time we were both transiting to the

[00:27:15] airport, his flight was delayed, of course, we're on the same flight. And rather him sort of saying, I'll see you on the other side and going into

[00:27:22] the lounge, right. And you know, maybe resting himself, you know, for the for the long flight ahead, he decided to sit with me in the in the in the

[00:27:29] regular area, right. And we were just talking and discussing life in general. And that had a profound effect on me, right, in terms of it doesn't

[00:27:35] matter how high you get up in any organization, understand that you are starting at some point in time at a at a much more singular individual

[00:27:41] contributor level. And it's important to have that level of engagement. So for me, Dr. Rise Nathan certainly will be a standout individual, mentor,

[00:27:48] teacher, and of course, somebody that I aspire to try to be like as well.

[00:27:52] Wow, what a powerful story. And one of the reasons I asked that question is, be probably blissfully unaware of how just investing that extra little

[00:28:00] bit of time with you sitting down and having a conversation, and how much that will inspire you and maybe inspire you to act the same when you rise up

[00:28:07] that ladder to incredibly powerful stuff. And for anyone listening that just wants to find out more information about anything we talked about

[00:28:14] today and explore SAP's data and analytics portfolio, obviously, SAP is a huge website. Is there anywhere in particular you'd like to point out

[00:28:21] everyone?

[00:28:23] So if you go across to the SAP.com, it's well structured, I think we've got a data foundation, the business technology platform will probably be the

[00:28:30] good starting point, which gives you the jumping point off into the data, the analytics, the planning space, in addition to the whole pro code, no

[00:28:37] code area as well. So a lot of related technology segments, we call it the business technology platform. And I think that will be a very good starting

[00:28:44] point.

[00:28:46] Well, it's been a huge pleasure talking with you today. We covered a lot there from gen AI, how it's powerless without contextual data and how keeping

[00:28:53] data clean for AI requires adapting to each specific data landscape, and also for helping business leaders understand how they can begin to practice better

[00:29:02] data hygiene for responsible AI use. So much gold in our conversation. As I said earlier in the episode there, there's an open invitation to get you back on

[00:29:10] and explore some of the other topics we flirted with today. But more than anything, thank you for joining me today.

[00:29:16] It's been a tremendous pleasure, Neil. Thank you so much.

[00:29:18] Wow, what a fascinating discussion there with Irfan. And I think his insights into SAP's data sphere, the importance of preserving data context and the ethical

[00:29:27] deployment of AI have all been incredibly enlightening. And Irfan's deep understanding of how data and AI can drive innovation across industries, I think provides a

[00:29:39] valuable perspective for everyone listening today. So I want to extend a special thank you to Irfan for his patience during our recording today as well, as some of you

[00:29:48] might know, I was having a new boiler fitted today, which meant I was using portable mics and connected to 5g due to having no electricity. But it was

[00:29:57] incredibly understanding and professionalism from today's guest that made the recording a success, despite some slightly unusual circumstances on my side.

[00:30:07] But as for everyone listening, I hope you enjoyed today's episode as much as I did. We'd love to hear from you your thoughts on today's discussion. What stood out to

[00:30:16] you about the future of data, the future of AI? Let me know your insights. Join the conversation by emailing me techblogwriteratlook.com, Twitter, LinkedIn,

[00:30:26] Instagram, just at Neil C Hughes. But remember, stay curious, stay innovative, keep exploring the limitless possibilities of technology. But more than

[00:30:38] anything, thank you for listening. And until next time, don't be a stranger.