1 00:00:24,440 --> 00:00:26,560 Music. Hello everyone, welcome to the 2 00:00:26,560 --> 00:00:29,680 next episode of the SAP BTP talk. 3 00:00:29,840 --> 00:00:32,240 My name is Nicholas. For everyone which don't already 4 00:00:32,240 --> 00:00:36,760 know me, I'm product manager for SAP BTP and I'm happy that I'm 5 00:00:36,760 --> 00:00:40,920 hosting this episode today. So the March episode, let's see 6 00:00:40,920 --> 00:00:42,760 what we have in the box for today. 7 00:00:42,760 --> 00:00:46,920 So first of all, as always, in our podcast, we have an extract 8 00:00:46,920 --> 00:00:50,360 of road map items. And then secondly, the main part 9 00:00:50,360 --> 00:00:55,400 will be, as always, an interview today having Shabana Samsuddin, 10 00:00:55,440 --> 00:00:58,120 Senior Product Manager for SAP HANA Cloud with me. 11 00:00:58,120 --> 00:01:02,680 And we will make a deep dive into the new knowledge graph 12 00:01:02,720 --> 00:01:07,800 engine of HANA Cloud. But let's start with the road 13 00:01:07,800 --> 00:01:11,360 map items for today. Please be aware that of course, 14 00:01:11,360 --> 00:01:14,720 road map items and road map information is under a 15 00:01:14,720 --> 00:01:17,320 disclaimer and is not guaranteed to be fulfilled. 16 00:01:19,720 --> 00:01:23,800 The first road map item of today is about SAP build work Zone. 17 00:01:23,880 --> 00:01:27,560 With the fulfillment of this road map item, generative AI 18 00:01:27,560 --> 00:01:32,760 capabilities will be introduced into build works in Advanced 19 00:01:32,760 --> 00:01:35,440 Edition. And with these AI capabilities, 20 00:01:35,440 --> 00:01:40,160 you are able to create, review and update your textual content 21 00:01:40,160 --> 00:01:44,320 in your work zone pages. This will make your process, of 22 00:01:44,320 --> 00:01:46,960 course, the process of the content creation and works in 23 00:01:46,960 --> 00:01:49,520 faster and more efficient for you as a tutor. 24 00:01:49,600 --> 00:01:52,840 So it can be used for the initial prompts all for 25 00:01:52,840 --> 00:01:56,520 reviewing and getting suggested on already written content. 26 00:01:57,360 --> 00:01:59,640 Next one is about SAP Integration Suite. 27 00:01:59,640 --> 00:02:03,720 So of course, integration is an essential part in modern IT and 28 00:02:03,720 --> 00:02:05,520 landscapes. We don't need to discuss about 29 00:02:05,520 --> 00:02:08,280 that. Therefore the topic of adapters 30 00:02:08,280 --> 00:02:11,320 is quite important for the cloud integration capability of the 31 00:02:11,320 --> 00:02:14,720 integration suite and in the future there will be a couple of 32 00:02:14,720 --> 00:02:18,000 new adapters available. I will give you 2 examples. 33 00:02:18,320 --> 00:02:23,080 The first one is an AI related adapter which provides you the 34 00:02:23,440 --> 00:02:27,680 optional capability to connect two large language models 35 00:02:27,680 --> 00:02:31,840 through SAP Integration Suite and the second one is an adapter 36 00:02:31,840 --> 00:02:37,080 for SAP Linux which allows you to connect to this SAP solution 37 00:02:37,200 --> 00:02:39,080 through the cloud integration capability. 38 00:02:40,800 --> 00:02:45,120 Third road map item of today is about SAP build apps and it's 39 00:02:45,120 --> 00:02:48,640 another AI road map item. Nowadays, SAP build apps will 40 00:02:48,640 --> 00:02:53,680 receive a feature that you can use AI to generate a user 41 00:02:53,680 --> 00:02:56,480 interface based on a given data source. 42 00:02:56,640 --> 00:02:59,840 So with this, you can boost your productivity or the productivity 43 00:02:59,840 --> 00:03:03,400 of your state in developers. So instead of creating a user 44 00:03:03,400 --> 00:03:06,480 interface from scratch, you can provide a data source. 45 00:03:06,680 --> 00:03:10,480 And yeah, I would say a skeleton of AUI will be generated. 46 00:03:11,320 --> 00:03:14,680 And the last format item for today's about SAP Process 47 00:03:14,680 --> 00:03:17,880 automation. Events are very crucial in a 48 00:03:17,880 --> 00:03:21,360 modern software architecture. And in the future, SAP build 49 00:03:21,360 --> 00:03:25,640 process automation can execute automations based on events from 50 00:03:25,640 --> 00:03:27,960 event brokers. So this will offer even more 51 00:03:27,960 --> 00:03:30,160 options to achieve an automated enterprise. 52 00:03:31,120 --> 00:03:34,920 As you can see a lot of new and helpful SAPBDP innovations 53 00:03:34,920 --> 00:03:36,720 planned for the upcoming weeks and months. 54 00:03:37,040 --> 00:03:39,400 Stay tuned. You can find more information 55 00:03:39,400 --> 00:03:42,760 and the links to the road mode items as always in the road 56 00:03:42,760 --> 00:03:47,240 notes or just go to rdmaps.sap.com and you'll find 57 00:03:47,240 --> 00:03:55,240 all product Rd. maps from SAP. And with that, I'm very happy to 58 00:03:55,240 --> 00:03:58,960 start today's interview with my guest, Shabana Samsudin. 59 00:03:59,000 --> 00:04:01,200 And yeah, warm welcome to our podcast. 60 00:04:01,200 --> 00:04:03,120 Shabana. It's a pleasure having you with 61 00:04:03,120 --> 00:04:05,720 me in this episode. Before we start into our 62 00:04:05,720 --> 00:04:09,760 exciting topic of Knowledge Graph Engine from Hammock Cloud, 63 00:04:10,040 --> 00:04:13,520 I would ask you to please first introduce yourself and give us 64 00:04:13,520 --> 00:04:17,519 some insights about you and your role at SAP as it's your first 65 00:04:17,519 --> 00:04:20,920 time in this podcast. Thank you so much, Nicholas. 66 00:04:20,959 --> 00:04:24,480 I'm really excited as well to be here and talk about the new 67 00:04:24,480 --> 00:04:27,920 knowledge graph engine and all the cool things that's happening 68 00:04:27,920 --> 00:04:31,320 in SAP HANA Cloud. To my introduction part, so I am 69 00:04:31,320 --> 00:04:35,280 a senior product manager at SAP working on the multi model 70 00:04:35,280 --> 00:04:38,720 engines of HANA Cloud. So that's including the spatial 71 00:04:38,720 --> 00:04:42,600 property graph, Jason document store vector and now the 72 00:04:42,600 --> 00:04:46,960 knowledge graph engine. My main focus here is basically 73 00:04:46,960 --> 00:04:51,280 helping customers adopt these technologies very effectively 74 00:04:51,280 --> 00:04:54,920 and efficiently on their side so that they can leverage the full 75 00:04:54,920 --> 00:04:58,760 power of SAP HANA Cloud for their diverse data needs. 76 00:04:59,360 --> 00:05:03,960 And before all this, I started my career as a QA test engineer, 77 00:05:03,960 --> 00:05:07,720 which gave me a very good foundation in understanding how 78 00:05:07,720 --> 00:05:12,360 systems work under the hood. So over time, I moved into 79 00:05:12,360 --> 00:05:15,920 product management where I now work closely with the customers 80 00:05:15,920 --> 00:05:20,280 and stakeholders, helping them, guiding them on how to integrate 81 00:05:20,280 --> 00:05:24,840 and get most out of these multi model capabilities of Hannah. 82 00:05:25,320 --> 00:05:28,320 That's all on the professional side and coming to the personal. 83 00:05:28,320 --> 00:05:33,840 And I'm a huge nature lover, so I love taking long walks, going 84 00:05:33,840 --> 00:05:37,680 on hikes, and just anything and everything about being outdoors. 85 00:05:37,680 --> 00:05:41,200 So when I'm not exploring, you'll probably find me reading 86 00:05:41,200 --> 00:05:44,120 a good book. So that's all about me, Natasha. 87 00:05:45,520 --> 00:05:46,640 Yeah. Thank you, Shabana, for 88 00:05:46,640 --> 00:05:50,360 introducing yourself so you find a good contrast between your 89 00:05:50,360 --> 00:05:52,560 professional and personal life. That's great. 90 00:05:53,000 --> 00:05:57,200 Yeah, last year March episode was about the HANA Vector 91 00:05:57,200 --> 00:05:59,920 engine. Today we're going to talk about 92 00:05:59,920 --> 00:06:02,560 the knowledge craft engine of SAP HANA Cloud. 93 00:06:02,840 --> 00:06:06,440 But let's start some steps back. If someone has never heard of a 94 00:06:06,440 --> 00:06:09,440 knowledge craft before, how would you explain it in simple 95 00:06:09,440 --> 00:06:11,760 terms? That's very much possible and 96 00:06:11,760 --> 00:06:15,440 that's a great question. I would say let's take an 97 00:06:15,440 --> 00:06:19,360 example to break all this down. OK, so imagine that you are 98 00:06:19,360 --> 00:06:23,600 planning a trip, let's say to Paris and first of all you would 99 00:06:24,040 --> 00:06:28,880 start by looking up on flights. Then you start checking out the 100 00:06:28,880 --> 00:06:31,840 hotels and then finding restaurants. 101 00:06:32,040 --> 00:06:35,560 Then maybe make a list of all the must visit places, right? 102 00:06:35,560 --> 00:06:40,120 So all of these things like the flights, hotels, landmarks, food 103 00:06:40,120 --> 00:06:44,480 spots are all connected in your mind because they all relate to 104 00:06:44,480 --> 00:06:46,800 your trip. So that's exactly what a 105 00:06:46,800 --> 00:06:50,880 knowledge graph does as well, but in a very structured and a 106 00:06:50,880 --> 00:06:54,120 digital way. Instead of storing information 107 00:06:54,120 --> 00:06:58,720 in traditional rows and columns like a spreadsheet, a knowledge 108 00:06:58,720 --> 00:07:02,600 graph basically connects data like a web, showing 109 00:07:02,600 --> 00:07:04,840 relationships between different entities. 110 00:07:05,840 --> 00:07:09,760 Again, for an example, in a business setting, it can be 111 00:07:09,760 --> 00:07:13,240 about linking customers, their transactions, it could be about 112 00:07:13,240 --> 00:07:15,360 linking products and their suppliers. 113 00:07:15,360 --> 00:07:18,720 So that's in a way that's much more intuitive. 114 00:07:19,280 --> 00:07:21,320 It's like giving your data a brain. 115 00:07:21,880 --> 00:07:26,520 It understands not just what things are but also how they are 116 00:07:26,520 --> 00:07:30,440 connected, making it much more easier to find like insights, 117 00:07:30,880 --> 00:07:34,040 come up with some design patterns or detect some of the 118 00:07:34,040 --> 00:07:38,880 patterns and also even power the AI applications very much needed 119 00:07:38,880 --> 00:07:42,400 in today's modern world. Basically in a nutshell, a 120 00:07:42,400 --> 00:07:45,680 knowledge graph would help businesses organize and connect 121 00:07:45,680 --> 00:07:50,840 their data in a much more natural as well as in a very 122 00:07:50,840 --> 00:07:53,600 intuitive way. That's I would just relate it 123 00:07:53,600 --> 00:07:56,640 back to how a human would think and process information. 124 00:07:57,800 --> 00:08:00,560 Hey, yeah, thank you. I have been to Paris last year, 125 00:08:00,560 --> 00:08:04,600 so maybe a knowledge graph had my planning be easier. 126 00:08:05,000 --> 00:08:10,520 But yeah anyways, let's have a look on the difference from a 127 00:08:10,520 --> 00:08:13,800 knowledge graph to a regular database of spreadsheet. 128 00:08:13,800 --> 00:08:17,920 You already introduced it slightly, but can you please get 129 00:08:17,920 --> 00:08:20,360 us more details? So what is really the difference 130 00:08:20,360 --> 00:08:23,240 between a knowledge graph and a regular database or a 131 00:08:23,240 --> 00:08:25,840 spreadsheet and default data type? 132 00:08:26,840 --> 00:08:29,840 I would say the biggest difference here is how the data 133 00:08:29,840 --> 00:08:34,720 is structured and connected. Oh, let's go back to the same 134 00:08:34,799 --> 00:08:36,840 example like the spreadsheet example. 135 00:08:36,840 --> 00:08:41,000 So it's like in a spreadsheet you have a list of data in rows 136 00:08:41,000 --> 00:08:43,559 and columns. If you want to find 137 00:08:43,559 --> 00:08:48,440 relationships you usually must look across multiple sheets or 138 00:08:48,440 --> 00:08:51,280 do some complicated look UPS. Also on the traditional 139 00:08:51,280 --> 00:08:55,080 databases and like the relational databases they are a 140 00:08:55,080 --> 00:08:59,800 bit more advanced but they still store data in form of tables but 141 00:08:59,800 --> 00:09:03,800 predefined relationships. Meaning they work well for 142 00:09:03,800 --> 00:09:07,480 structured data, but they eventually could struggle with 143 00:09:07,480 --> 00:09:09,880 complex and interconnected information. 144 00:09:10,800 --> 00:09:14,360 And now a knowledge graph takes a completely different approach 145 00:09:14,360 --> 00:09:17,520 here. So instead of just storing data 146 00:09:17,520 --> 00:09:22,360 in a very rigid format, it starts to connect data points in 147 00:09:22,360 --> 00:09:25,680 a web like structure. So you could imagine instead of 148 00:09:25,680 --> 00:09:29,480 having separate tables for customers and orders, you have a 149 00:09:29,480 --> 00:09:33,960 graph that a customer is directly linked to their orders, 150 00:09:34,560 --> 00:09:37,640 which are then linked to products, which are then linked 151 00:09:37,640 --> 00:09:41,160 to suppliers. So this makes it incredibly easy 152 00:09:41,160 --> 00:09:45,640 to navigate relationships and also come across or uncover some 153 00:09:45,640 --> 00:09:49,400 of the hidden connections here. And basically again, powering 154 00:09:49,400 --> 00:09:54,640 the AI insights that's needed. So if a traditional database is 155 00:09:54,640 --> 00:09:58,800 like file cabinet where everything is stored in separate 156 00:09:58,800 --> 00:10:02,360 folders, a knowledge graph is more like a mind map, like 157 00:10:02,360 --> 00:10:05,680 basically showing how everything is connected in real time. 158 00:10:05,920 --> 00:10:09,240 It will eventually make the search and an analysis much more 159 00:10:09,240 --> 00:10:13,160 intuitive and easier. Yeah, thank you for explaining 160 00:10:13,160 --> 00:10:15,400 that. I like the idea of a mind map 161 00:10:15,400 --> 00:10:18,680 more than having a VAP because most of the people don't like 162 00:10:18,680 --> 00:10:22,640 spiralus. But OK, let's get into the 163 00:10:22,840 --> 00:10:25,000 examples. You already had one example with 164 00:10:25,000 --> 00:10:29,000 the Paris strip, but can you give us some more real world 165 00:10:29,000 --> 00:10:31,120 examples of how knowledge graphs are used? 166 00:10:31,960 --> 00:10:36,240 Of course. So knowledge graphs as we speak 167 00:10:36,240 --> 00:10:40,000 are already powering so many things that we use on a daily 168 00:10:40,000 --> 00:10:44,360 basis, sometimes even without us realizing it. 169 00:10:44,640 --> 00:10:48,320 So let me take some real world examples, like few of them. 170 00:10:49,080 --> 00:10:54,120 First and foremost, Google search, what we usually do, we 171 00:10:54,120 --> 00:10:58,160 eventually think about going and searching for a celebrity or a 172 00:10:58,160 --> 00:11:02,280 landmark or a movie, all these sort of things that we want to 173 00:11:02,440 --> 00:11:05,720 Google search, right? So when you search about these, 174 00:11:05,960 --> 00:11:10,920 you get a very neat summary of the search item with every 175 00:11:10,920 --> 00:11:14,280 related facts and images and links, right? 176 00:11:14,720 --> 00:11:17,200 That's exactly Knowledge Graph at work there. 177 00:11:17,680 --> 00:11:21,560 Instead of just showing the web pages, it's kind of already 178 00:11:21,560 --> 00:11:25,160 connecting the related information for you so that you 179 00:11:25,160 --> 00:11:27,240 get the answers very instantaneously. 180 00:11:27,240 --> 00:11:31,000 But also it's much easier for you to understand how everything 181 00:11:31,000 --> 00:11:32,880 is relating to your search object. 182 00:11:33,840 --> 00:11:37,960 Another interesting example would be the fraud deduction in 183 00:11:37,960 --> 00:11:41,160 banks, right? So banks use knowledge graphs to 184 00:11:41,160 --> 00:11:43,720 detect unusual financial patterns. 185 00:11:44,080 --> 00:11:48,000 Let's say someone is suddenly starting to make multiple large 186 00:11:48,000 --> 00:11:51,280 transactions or transfers to a high risk account. 187 00:11:51,280 --> 00:11:55,360 Then a graph can help you map out those suspicious connections 188 00:11:55,480 --> 00:11:59,000 and also help you in flagging potential fraud before it 189 00:11:59,000 --> 00:12:02,800 eventually happens. And also another very relatable 190 00:12:02,800 --> 00:12:05,600 example I could say is the Netflix and Spotify 191 00:12:05,600 --> 00:12:08,320 recommendations. When you have your Netflix 192 00:12:08,320 --> 00:12:12,640 account, it basically suggests shows based on what you have 193 00:12:12,640 --> 00:12:17,480 watched previously or Spotify which creates a custom playlist 194 00:12:17,480 --> 00:12:19,600 for you. So they are using knowledge 195 00:12:19,600 --> 00:12:22,280 graphs. It is linking the users and 196 00:12:22,280 --> 00:12:26,400 their preferences and their content in a much smarter way to 197 00:12:26,400 --> 00:12:29,560 make the recommendations feel even more personal. 198 00:12:30,080 --> 00:12:34,360 And I will end it with business use case, let's say supply chain 199 00:12:34,360 --> 00:12:36,800 and logistics. So businesses are using 200 00:12:36,800 --> 00:12:40,640 knowledge graphs to track suppliers, their shipments and 201 00:12:40,640 --> 00:12:44,640 inventories, all of this. So if one supplier, say for 202 00:12:44,640 --> 00:12:48,640 example, has a delay, the system can then quickly find 203 00:12:48,680 --> 00:12:51,840 alternatives based on the existing relationships. 204 00:12:52,040 --> 00:12:55,000 And also it is definitely helping the company stay ahead 205 00:12:55,000 --> 00:12:58,160 of disruptions in a much more intuitive way again. 206 00:13:00,720 --> 00:13:02,280 Yeah, yeah. Thank you for sharing these 207 00:13:02,280 --> 00:13:04,160 examples. And I'm very happy that 208 00:13:04,160 --> 00:13:06,200 knowledge grabs are protecting their banking. 209 00:13:06,720 --> 00:13:10,920 But in general, why should we care about knowledge grabs? 210 00:13:10,920 --> 00:13:14,120 So why do they matter? I mean, we have these examples 211 00:13:14,120 --> 00:13:17,320 now, but yeah, let's get more into detail. 212 00:13:17,920 --> 00:13:19,960 Yeah, that's a very interesting question. 213 00:13:20,200 --> 00:13:22,280 Why should we care about knowledge grab? 214 00:13:23,040 --> 00:13:27,120 The short answer here would be because they would help us make 215 00:13:27,120 --> 00:13:31,880 sense of complex information in a way that traditional databases 216 00:13:31,880 --> 00:13:35,400 simply can't. So let's be real, in today's 217 00:13:35,400 --> 00:13:38,200 world, data is really exploding, right? 218 00:13:38,520 --> 00:13:42,200 Every company, every industry is drowning with data. 219 00:13:42,480 --> 00:13:46,840 But withdraw data alone, it doesn't make much sense unless 220 00:13:46,840 --> 00:13:49,800 we can understand how much things are connected. 221 00:13:50,320 --> 00:13:53,120 There's where I believe this knowledge crafts could come into 222 00:13:53,120 --> 00:13:56,080 picture. So some bullet points on why 223 00:13:56,080 --> 00:13:59,560 knowledge crafts really matter. They basically help you in 224 00:13:59,560 --> 00:14:03,400 turning your data into insights. Like I said, instead of just 225 00:14:03,400 --> 00:14:06,880 storing your information on knowledge graph is helping you 226 00:14:07,200 --> 00:14:10,680 discover some hidden patterns and relationships and trends. 227 00:14:10,960 --> 00:14:14,040 It could be detecting fraud or improving search or even 228 00:14:14,040 --> 00:14:18,080 predicting supply chain risks. They're also making the AI and 229 00:14:18,080 --> 00:14:20,960 machine learning smarter AI models. 230 00:14:20,960 --> 00:14:25,720 They need context to make better predictions and knowledge graphs 231 00:14:25,720 --> 00:14:30,400 provide that by linking the data points in a way that mimics 232 00:14:30,400 --> 00:14:33,280 human thinking. So This is why they are becoming 233 00:14:33,280 --> 00:14:36,920 essential for AI driven applications like for example 234 00:14:36,920 --> 00:14:39,960 chat bots or recommendation systems and even intelligence 235 00:14:39,960 --> 00:14:43,520 systems. Most importantly, they break 236 00:14:43,520 --> 00:14:47,800 down the data silos I would say because businesses often have 237 00:14:47,800 --> 00:14:51,760 data that is scattered across multiple systems, like it could 238 00:14:51,760 --> 00:14:55,800 be customer databases, financial records or supply chain logs. 239 00:14:56,280 --> 00:14:59,160 So the knowledge graph could help in connecting the dot 240 00:14:59,160 --> 00:15:02,920 between all of these sources, making it much easier and also 241 00:15:02,920 --> 00:15:06,440 giving you a very full picture of your business scenarios. 242 00:15:07,360 --> 00:15:10,480 Last but not the least, they save time and they reduce 243 00:15:10,480 --> 00:15:13,600 complexity. So let's say you are a customer 244 00:15:13,600 --> 00:15:17,120 service or you are working in the customer service field and 245 00:15:17,120 --> 00:15:21,160 you need to understand a customer's history, what they 246 00:15:21,160 --> 00:15:24,200 have bought in the past, what are their past support tickets 247 00:15:24,200 --> 00:15:26,400 or their feedback on your products and so on. 248 00:15:26,640 --> 00:15:30,200 Instead of just searching across multiple systems, a knowledge 249 00:15:30,200 --> 00:15:33,080 graph could help you in connecting all the dots and also 250 00:15:33,080 --> 00:15:36,880 giving you a very instant view of everything that's related to 251 00:15:36,880 --> 00:15:39,080 the customers. Basically tying it back to the 252 00:15:39,080 --> 00:15:41,640 Google search example that I gave in the beginning. 253 00:15:42,600 --> 00:15:47,840 Making AI smarter, breaking down data silos, reducing complexity. 254 00:15:47,960 --> 00:15:51,840 OK, Yeah, I'm with you. We should care about knowledge 255 00:15:51,840 --> 00:15:55,160 graphs. And the good thing is that 256 00:15:55,440 --> 00:15:58,720 there's a brand new knowledge graph engine in SAP HANA Cloud 257 00:15:58,720 --> 00:16:00,280 now. So what does it mean for the 258 00:16:00,280 --> 00:16:02,200 customers? What are the benefits for 259 00:16:02,200 --> 00:16:05,520 customers of HANA Cloud? That comes again the interesting 260 00:16:05,520 --> 00:16:07,200 part. Very great question. 261 00:16:07,200 --> 00:16:10,720 Thanks Nicholas for that. So the addition of knowledge 262 00:16:10,720 --> 00:16:15,400 graphs to SAP HANA Cloud is a big deal because it is going to 263 00:16:15,400 --> 00:16:19,840 give customers a completely new way to connect and analyse their 264 00:16:19,840 --> 00:16:22,320 data. Like in the sense I have already 265 00:16:22,320 --> 00:16:25,720 mentioned it before, rather than just storing it, you will have 266 00:16:25,720 --> 00:16:29,240 an intuitive way to connect your data and also uncover the 267 00:16:29,240 --> 00:16:31,880 insights alongside the relational data that you're 268 00:16:31,880 --> 00:16:35,840 going to store on HANA Cloud. Businesses today just they don't 269 00:16:36,000 --> 00:16:39,680 deal just with the isolated data points, right? 270 00:16:39,680 --> 00:16:42,960 So they need an understanding and relationship between these 271 00:16:42,960 --> 00:16:45,320 data O. That is what Knowledge Grah and 272 00:16:45,320 --> 00:16:47,640 the Knowledge Grah engine is going to enable at the end of 273 00:16:47,640 --> 00:16:50,880 the day. For customers, this means that 274 00:16:50,880 --> 00:16:54,520 they can now move beyond the traditional database queries and 275 00:16:54,520 --> 00:16:58,280 start discovering these hidden connections within their data. 276 00:16:58,520 --> 00:17:02,320 Instead of running complex joins across multiple tables, they can 277 00:17:02,360 --> 00:17:06,200 instantly see relationships, whether that's between, again, 278 00:17:06,200 --> 00:17:10,040 customer suppliers, financial transactions or even products. 279 00:17:10,200 --> 00:17:13,560 So this is incredibly valuable in the areas like fraud 280 00:17:13,560 --> 00:17:17,760 detection, customer analytics, and also just basically 281 00:17:17,760 --> 00:17:20,000 understanding relationships in real time. 282 00:17:20,400 --> 00:17:23,000 It definitely does make a lot of difference for them. 283 00:17:23,920 --> 00:17:29,040 Another key benefit I would say is how this basically is going 284 00:17:29,040 --> 00:17:31,480 to improve the AI driven applications. 285 00:17:31,880 --> 00:17:36,240 So many businesses nowadays are slowly starting to rely on AI 286 00:17:36,240 --> 00:17:38,680 and machine learning for automation and making 287 00:17:38,680 --> 00:17:42,760 predictions and recommendations. Right Knowledge Graph would 288 00:17:42,760 --> 00:17:47,000 strengthen these capabilities by providing much deeper context 289 00:17:47,000 --> 00:17:51,320 and also helping the AIS make some better decisions with this 290 00:17:51,320 --> 00:17:54,520 richer and connected data that the customers can now store. 291 00:17:55,520 --> 00:17:59,920 And finally, the best part, it all happens in SAP HANA Cloud. 292 00:17:59,920 --> 00:18:04,640 So businesses need not manage separate tools or move around 293 00:18:04,640 --> 00:18:06,920 their data between different systems. 294 00:18:07,000 --> 00:18:11,440 That means they can just now store, process and analyse these 295 00:18:11,440 --> 00:18:16,160 highly connected Knowledge graph data very much efficiently, all 296 00:18:16,160 --> 00:18:19,480 within a single database. That's like in the same 297 00:18:19,480 --> 00:18:22,720 environment. Yeah, thanks for sharing. 298 00:18:22,720 --> 00:18:27,480 And I mean, we already had the term in your introduction, the 299 00:18:27,480 --> 00:18:32,120 multi model engine or the multi model data processing from SAP 300 00:18:32,160 --> 00:18:34,120 HANA Cloud. So the knowledge graph engine is 301 00:18:34,120 --> 00:18:37,480 one part of it for everyone which never heard about it in 302 00:18:37,480 --> 00:18:39,920 the past. Can you summarize what the multi 303 00:18:39,960 --> 00:18:44,320 model data processing means? Or Yeah what's in the box for 304 00:18:44,320 --> 00:18:46,000 the customers? Absolutely. 305 00:18:46,240 --> 00:18:49,960 Like I said, I'm an evangelist for multi model processing in 306 00:18:49,960 --> 00:18:52,920 HANA Cloud. Coming to explaining it, multi 307 00:18:52,920 --> 00:18:57,920 model data processing might just sound very technical, but it's 308 00:18:57,920 --> 00:19:03,160 really about making life and the processes easier for businesses 309 00:19:03,160 --> 00:19:05,920 that are dealing with different types of data. 310 00:19:05,920 --> 00:19:10,920 So traditionally companies store data in different databases 311 00:19:11,160 --> 00:19:15,320 depending on the type that is structured data in relational 312 00:19:15,320 --> 00:19:19,640 databases, they could be storing documents in no SQL stores or 313 00:19:19,640 --> 00:19:23,600 graph data in separate graph databases and then obviously the 314 00:19:23,600 --> 00:19:27,240 spatial data in geographic information systems and so on. 315 00:19:27,760 --> 00:19:30,000 So there is a problem here, right? 316 00:19:30,000 --> 00:19:34,120 So these systems don't always talk to each other easily. 317 00:19:34,360 --> 00:19:36,840 This could be leading to data silos. 318 00:19:36,840 --> 00:19:42,520 A lot of extra effort to combine insights across multiple systems 319 00:19:42,520 --> 00:19:46,200 that are involved here. But with the HANA Clouds multi 320 00:19:46,200 --> 00:19:50,800 model processing, everything is handled in a single and unified 321 00:19:50,800 --> 00:19:54,400 database as I mentioned before. So that means businesses can 322 00:19:54,400 --> 00:19:59,720 store and analyse relational document kind of data or graph 323 00:19:59,720 --> 00:20:04,040 data, spatial and even time series data, all of them in one 324 00:20:04,040 --> 00:20:06,640 place. So there's no need to move data 325 00:20:06,640 --> 00:20:10,200 between different systems or even worry about performances. 326 00:20:10,200 --> 00:20:14,120 The trade-offs that are involved here, it is all going to work 327 00:20:14,120 --> 00:20:16,640 together very seamlessly for the customers. 328 00:20:17,200 --> 00:20:19,960 I would like to explain it better with an example. 329 00:20:20,160 --> 00:20:24,960 Let's imagine like we have a retail company that's managing 330 00:20:24,960 --> 00:20:28,160 its supply chain. So they might be storing 331 00:20:28,160 --> 00:20:31,120 transaction records in relational tables. 332 00:20:31,360 --> 00:20:34,440 They need to track customer interactions that could be 333 00:20:34,440 --> 00:20:38,840 stored as Jason documents and they could be using graph models 334 00:20:38,840 --> 00:20:42,520 to understand the relationships between the suppliers and they 335 00:20:42,520 --> 00:20:47,240 could be relying on spatial data to understand and optimize the 336 00:20:47,240 --> 00:20:51,080 delivery routes. So instead of juggling between 337 00:20:51,080 --> 00:20:54,840 multiple databases here, they could rely on SAP HANA Cloud 338 00:20:54,840 --> 00:20:59,080 which is allowing them to query and analyse all of these data 339 00:20:59,080 --> 00:21:04,120 types together in a much faster way, which in the sense is also 340 00:21:04,120 --> 00:21:07,040 providing them better decision making capabilities. 341 00:21:08,200 --> 00:21:10,440 Also. In short, multi model processing 342 00:21:10,440 --> 00:21:14,080 is about providing a lot of flexibility and efficiency for 343 00:21:14,080 --> 00:21:16,360 our customers. So it is eliminating the 344 00:21:16,360 --> 00:21:20,600 complexity for them in the sense that using the headache of doing 345 00:21:20,600 --> 00:21:24,560 a lot of integrations and ETL kind of transfers between 346 00:21:24,560 --> 00:21:28,160 multiple systems. And it is also making it much 347 00:21:28,160 --> 00:21:31,440 easier for them to extract some meaningful insights out of their 348 00:21:31,440 --> 00:21:33,800 data. No matter how it is structured, 349 00:21:34,000 --> 00:21:36,520 it could be structured and structured or semi structured. 350 00:21:36,960 --> 00:21:39,960 And now with the Knowledge Graph engine as part of HANA Cloud, 351 00:21:39,960 --> 00:21:44,080 businesses can definitely take their data analysis one step 352 00:21:44,080 --> 00:21:46,640 further. Like we have been talking since 353 00:21:46,640 --> 00:21:49,400 the beginning, just understanding the relationships 354 00:21:49,400 --> 00:21:52,040 and connections in a whole new different way. 355 00:21:53,040 --> 00:21:56,840 Thanks Urbana for summarizing and explaining what multi model 356 00:21:56,840 --> 00:21:59,560 data processing means once again for our listeners. 357 00:21:59,840 --> 00:22:03,760 And I believe this was a very good explanation and also a very 358 00:22:03,760 --> 00:22:06,600 good presentation of the business value of SAP on the 359 00:22:06,600 --> 00:22:09,440 cloud for everyone, which isn't a customer so far, right? 360 00:22:09,440 --> 00:22:12,600 So reducing the complexity and bringing all the different data 361 00:22:12,600 --> 00:22:15,000 types together in one place sounds great. 362 00:22:17,480 --> 00:22:21,280 I would say let's refresh our minds with a small power break 363 00:22:21,280 --> 00:22:23,040 game. I'm giving you three small 364 00:22:23,040 --> 00:22:27,080 questions like coffee or energy drink and you just tell us a 365 00:22:27,080 --> 00:22:28,760 choice. Do you want to play this Power 366 00:22:28,760 --> 00:22:32,000 Break game with me? Yeah, let's be it sounds very 367 00:22:32,000 --> 00:22:34,480 fun. Let's get into it before getting 368 00:22:34,480 --> 00:22:36,280 into much more details of Knowledge Graph. 369 00:22:36,560 --> 00:22:39,160 Exactly. So then the first question would 370 00:22:39,160 --> 00:22:42,920 be, do you prefer swimming in the ocean or swimming in the 371 00:22:42,920 --> 00:22:45,440 pool? So very funny question. 372 00:22:46,160 --> 00:22:49,280 I think swimming in Pool, I would say I'm a little bit of a 373 00:22:49,280 --> 00:22:52,760 terrible swimmer, so I can't take the risk at the ocean. 374 00:22:53,040 --> 00:22:55,200 So pool sounds a little safer for me. 375 00:22:56,120 --> 00:22:58,640 Sounds fair. Then the second question, do you 376 00:22:58,640 --> 00:23:03,800 prefer sunny days or rainy days? Sunny days all the way for the 377 00:23:03,800 --> 00:23:06,360 wind. I just hate when the weather is 378 00:23:06,360 --> 00:23:09,520 so gloomy. So I think sunny and bright is 379 00:23:09,520 --> 00:23:11,400 my style. Same for me. 380 00:23:11,400 --> 00:23:14,200 And then the third question, unfortunately, I already know 381 00:23:14,200 --> 00:23:16,440 the answer. I should have prepared another 382 00:23:16,440 --> 00:23:19,080 question because you already introduced it in yourself. 383 00:23:19,080 --> 00:23:21,000 Introduction. But watching a film or reading a 384 00:23:21,000 --> 00:23:25,360 book? OK, let's give it a small twist 385 00:23:25,360 --> 00:23:26,800 here. I would say 5050. 386 00:23:27,000 --> 00:23:31,120 I'm also much of a movie lover. It basically would depend on my 387 00:23:31,120 --> 00:23:34,280 mood. If I want to be calm, quiet and 388 00:23:34,280 --> 00:23:36,720 just gain more knowledge, I would turn to books. 389 00:23:36,720 --> 00:23:40,520 Or if I just want to unwind and relax, then maybe a movie. 390 00:23:41,280 --> 00:23:44,520 Down here. OK, so thank you for playing the 391 00:23:44,520 --> 00:23:47,600 Power Break game with me to refresh our minds and especially 392 00:23:47,600 --> 00:23:53,080 the minds of our listeners. Coming back to the topic, so 393 00:23:53,080 --> 00:23:57,440 getting our hands dirty as much as it's possible in a podcast, 394 00:23:58,000 --> 00:24:02,440 how do I create knowledge graphs in general and how in SAP Hunter 395 00:24:02,440 --> 00:24:05,400 Cloud? Very, very interesting question. 396 00:24:05,400 --> 00:24:08,800 And also I believe this will be a question that many of our 397 00:24:08,800 --> 00:24:10,440 customers will have on their mind. 398 00:24:10,880 --> 00:24:14,040 So I'm happy to just step in there and answer that. 399 00:24:14,600 --> 00:24:18,640 Like in general, creating knowledge graph is basically 400 00:24:18,640 --> 00:24:22,040 involving like you have to define the entities, for 401 00:24:22,040 --> 00:24:25,080 example, the customers, the products and transactions 402 00:24:25,080 --> 00:24:27,760 involved, and then the relationships between them, 403 00:24:28,000 --> 00:24:30,120 right? So basically forming a 404 00:24:30,120 --> 00:24:32,560 structured and connected web of data. 405 00:24:32,920 --> 00:24:37,360 The goal here is to move beyond traditional tables and instead 406 00:24:37,360 --> 00:24:41,320 of that start to capture how different pieces of information 407 00:24:41,320 --> 00:24:43,680 are linked. It's not going to be very 408 00:24:43,680 --> 00:24:46,360 traditional like we have been discussing since the beginning, 409 00:24:46,880 --> 00:24:51,880 and in SAP HANA Cloud there is no built in way to automatically 410 00:24:51,880 --> 00:24:54,120 generate your knowledge graph at this point. 411 00:24:54,520 --> 00:24:59,560 Instead, customers are expected to model and build their own 412 00:24:59,560 --> 00:25:03,160 knowledge graphs based on their specific data and business 413 00:25:03,160 --> 00:25:07,680 needs. However, HANA Cloud is going to 414 00:25:07,680 --> 00:25:11,080 provide the knowledge graph processing capabilities that is 415 00:25:11,080 --> 00:25:14,040 needed to like you can have your own knowledge graph that's 416 00:25:14,040 --> 00:25:16,640 created on your side. You can efficiently store it, 417 00:25:16,640 --> 00:25:20,880 query it and also analyse the knowledge graphs alongside the 418 00:25:20,880 --> 00:25:24,880 other relational or unstructured kind of data that could be 419 00:25:24,880 --> 00:25:27,680 sitting on HANA Cloud. But in the sense knowledge 420 00:25:27,680 --> 00:25:31,440 graphs is again providing you a very structured way of doing it. 421 00:25:32,640 --> 00:25:36,240 And for those who are looking for easier ways to generate 422 00:25:36,240 --> 00:25:38,800 knowledge graphs, I would like to give some tips here. 423 00:25:38,880 --> 00:25:42,160 So they could be looking at some emerging tools that that 424 00:25:42,160 --> 00:25:45,720 basically help in automating the process of creation of these 425 00:25:45,720 --> 00:25:49,160 knowledge graphs. Like for example, one of the one 426 00:25:49,160 --> 00:25:51,920 of the things that I've been seeing very frequently at this 427 00:25:51,920 --> 00:25:55,280 point is the LLM based approaches, like for example the 428 00:25:55,280 --> 00:25:59,360 Lang chains graph transformer, where a large language model can 429 00:25:59,360 --> 00:26:04,400 process the unstructured text and it helps in extracting the 430 00:26:04,400 --> 00:26:07,480 structured entities and relationships out of this. 431 00:26:07,640 --> 00:26:12,240 So this is basically very useful if you have a lot of textual 432 00:26:12,240 --> 00:26:15,360 data, like you could be dealing with customer reviews or 433 00:26:15,360 --> 00:26:19,160 documents or even research papers and then you want to 434 00:26:19,160 --> 00:26:21,520 automatically build a knowledge graph out of it. 435 00:26:21,680 --> 00:26:23,720 And then this might be a way for you. 436 00:26:24,040 --> 00:26:27,240 And there are also multiple other approaches in building 437 00:26:27,240 --> 00:26:31,160 pre-existing ontologies or like taking advantage of what is 438 00:26:31,160 --> 00:26:35,840 already existing or some schema based knowledge extraction tools 439 00:26:35,840 --> 00:26:39,600 that could help you structure your data into a graph format, 440 00:26:39,600 --> 00:26:42,920 which can then be imported into HANA Cloud for creating an 441 00:26:42,920 --> 00:26:45,400 analysis as part of the knowledge graph engine. 442 00:26:45,760 --> 00:26:50,720 So just summarizing all that back, HANA Cloud does not at the 443 00:26:50,720 --> 00:26:55,400 moment provide an automated way to generate the knowledge graph, 444 00:26:55,400 --> 00:26:59,040 but it is providing a high performance engine that's needed 445 00:26:59,040 --> 00:27:01,720 to manage and analyze them efficiently. 446 00:27:02,320 --> 00:27:05,880 Thanks for explaining how to create knowledge graphs so as we 447 00:27:05,880 --> 00:27:09,080 know know how to create and store them inside of HANA Cloud. 448 00:27:09,200 --> 00:27:13,920 This knowledge graphs within SAP Anna Cloud have an impact on the 449 00:27:13,920 --> 00:27:15,920 day-to-day operations of the database. 450 00:27:17,040 --> 00:27:21,680 Again, a very great question. So one of the biggest advantages 451 00:27:21,680 --> 00:27:26,200 of using the knowledge graph engine within HANA Cloud is how 452 00:27:26,200 --> 00:27:29,880 easy it is going to be for you to integrate your data into the 453 00:27:29,880 --> 00:27:31,920 database. Meaning you don't have to worry 454 00:27:31,920 --> 00:27:35,360 about managing separate systems or dealing with any performance 455 00:27:35,360 --> 00:27:38,560 bottlenecks as a result of juggling with multiple systems, 456 00:27:38,560 --> 00:27:41,360 right? Because HANA Cloud is designed 457 00:27:41,360 --> 00:27:44,600 for multi model processing, so it is efficiently handling the 458 00:27:44,600 --> 00:27:47,960 knowledge graph data also alongside your relational 459 00:27:47,960 --> 00:27:51,320 document and spatial data in a native way. 460 00:27:51,840 --> 00:27:55,320 In fact, in many cases, using a knowledge graph can enhance the 461 00:27:55,320 --> 00:27:58,880 operations because it's going to allow the businesses to run much 462 00:27:58,880 --> 00:28:03,120 more faster than intuitively and also run some intuitive queries 463 00:28:03,120 --> 00:28:06,400 without having to do some complex joints between multiple 464 00:28:06,400 --> 00:28:09,040 tables. For example, like instead of 465 00:28:09,040 --> 00:28:13,520 executing multiple nested SQL queries to trace relationships 466 00:28:13,560 --> 00:28:17,600 in a traditional relational database, a graph query can 467 00:28:17,600 --> 00:28:21,600 retrieve those connections in just a couple of steps or even I 468 00:28:21,600 --> 00:28:25,080 would say just in a single step, which is going to significantly 469 00:28:25,080 --> 00:28:28,720 improve the efficiency. And since it's all running in 470 00:28:28,720 --> 00:28:32,600 HANA in memory database, the businesses tend to get the high 471 00:28:32,600 --> 00:28:36,600 performance and real time insights, providing the database 472 00:28:36,720 --> 00:28:39,640 with very minimal operational overhead on their side. 473 00:28:40,680 --> 00:28:43,120 OK. So as in summary, that means no 474 00:28:43,120 --> 00:28:46,200 additional day-to-day impact on the operations, right? 475 00:28:46,680 --> 00:28:50,280 Yeah, absolutely no. Nice, thanks for explaining. 476 00:28:50,280 --> 00:28:53,680 So then a very important question, does the Knowledge 477 00:28:53,680 --> 00:28:56,800 Graph engine of SAP HANA Cloud has an additional price tag for 478 00:28:56,800 --> 00:28:59,800 the customers? Yeah, this is also going to be 479 00:28:59,800 --> 00:29:02,240 the question on lot of customers mind. 480 00:29:02,240 --> 00:29:06,560 And here I have a good news. There's no extra cost to use the 481 00:29:06,560 --> 00:29:09,440 knowledge craft capabilities in SAP HANA Cloud. 482 00:29:09,440 --> 00:29:13,480 So it's going to be included as part of the overall HANA Cloud 483 00:29:13,480 --> 00:29:16,520 offering. So if you're already using SAP 484 00:29:16,560 --> 00:29:19,280 HANA Cloud, you can start leveraging knowledge crafts 485 00:29:19,280 --> 00:29:22,400 right away without having to deal with any additional 486 00:29:22,400 --> 00:29:26,440 licensing fees. And of course, like any other 487 00:29:26,440 --> 00:29:30,200 data features, the cost will also depend on how much of data 488 00:29:30,200 --> 00:29:34,440 you're processing and storing as well as also on the compute 489 00:29:34,440 --> 00:29:36,400 resources that you're going to be using. 490 00:29:36,680 --> 00:29:40,760 So since HANA Cloud is operating on consumption based pricing 491 00:29:40,760 --> 00:29:44,440 model, the customers can pay for the actual storage and the 492 00:29:44,440 --> 00:29:47,200 processing and the queries that they are running on their side. 493 00:29:47,400 --> 00:29:50,840 So this is just like they would for any other workload in HANA 494 00:29:50,840 --> 00:29:52,240 Cloud. It's not going to be any 495 00:29:52,240 --> 00:29:56,120 different here. So yeah, just in short, no extra 496 00:29:56,120 --> 00:30:00,800 license needed, no hidden fees, just a powerful new capability 497 00:30:00,800 --> 00:30:05,440 that the customers can start using immediately to basically 498 00:30:05,440 --> 00:30:07,800 unlock some deeper insights from their data. 499 00:30:08,360 --> 00:30:11,680 So I believe that sounds like a pretty good deal, right, 500 00:30:11,680 --> 00:30:14,480 Nicholas? Yeah, it sounds like a pretty 501 00:30:14,480 --> 00:30:17,720 good deal and good news for existing SAP HANA Cloud 502 00:30:17,720 --> 00:30:20,320 customers. All right, so let's get to the 503 00:30:20,320 --> 00:30:23,120 next question then. How do the knowledge graphs 504 00:30:23,200 --> 00:30:27,960 enhance AI models, particularly in search engines, 505 00:30:27,960 --> 00:30:30,920 recommendation systems, large language models? 506 00:30:31,280 --> 00:30:35,200 And can Knowledge Graph of SAP HANA Cloud be used in the same 507 00:30:35,200 --> 00:30:38,120 way? That's a fantastic question 508 00:30:38,240 --> 00:30:41,720 because I have been saying since the beginning, knowledge graphs 509 00:30:41,720 --> 00:30:46,320 and AI are a perfect match. So coming to the AI models, 510 00:30:46,320 --> 00:30:50,640 especially in search engines like recommendation systems, 511 00:30:50,640 --> 00:30:55,280 large language models, they work the best when they have context. 512 00:30:55,640 --> 00:30:58,440 And that's exactly what the knowledge graph is going to 513 00:30:58,440 --> 00:31:01,840 provide. So the knowledge graphs is going 514 00:31:01,840 --> 00:31:06,800 to structure data in a way that would help AI understand not 515 00:31:06,800 --> 00:31:11,280 just isolated facts, but also help them understand how things 516 00:31:11,280 --> 00:31:14,720 are related. Let's take the search engines 517 00:31:14,720 --> 00:31:17,880 like Google as an example. When you search for something 518 00:31:18,000 --> 00:31:21,880 like Leonardo da Vinci, you don't just get a list of web 519 00:31:22,040 --> 00:31:25,600 pages, right? You get an info box showing his 520 00:31:25,600 --> 00:31:30,320 works, his lifetime, his timeline of doing things, and 521 00:31:30,320 --> 00:31:32,880 all the related figures to his personal life. 522 00:31:33,080 --> 00:31:36,400 So that's powered by knowledge graph, which is enhancing the AI 523 00:31:36,400 --> 00:31:40,360 models understanding of entities and relationships for the search 524 00:31:40,360 --> 00:31:42,760 object. And then there's large language 525 00:31:42,760 --> 00:31:46,480 models like the ChatGPT. It's pretty famous nowadays. 526 00:31:46,480 --> 00:31:51,360 So while LLMS are great at generating text, they can 527 00:31:51,360 --> 00:31:55,120 sometimes produce misleading or hallucinating responses. 528 00:31:55,560 --> 00:32:00,200 When combined with knowledge graphs, the AI is able to get a 529 00:32:00,200 --> 00:32:03,800 verified and structured source of truth, which is going to 530 00:32:03,800 --> 00:32:07,720 improve their accuracy, especially for enterprise AI 531 00:32:07,720 --> 00:32:11,800 applications in the area of finance, healthcare or legal 532 00:32:11,800 --> 00:32:13,600 industries. Very much important. 533 00:32:14,520 --> 00:32:19,400 And can kind of clouds knowledge graph engine be used in the same 534 00:32:19,400 --> 00:32:20,840 way? Absolutely. 535 00:32:20,840 --> 00:32:25,720 Since it's integrated into the HANA Cloud ecosystem, businesses 536 00:32:25,720 --> 00:32:29,640 can use it to enhance their AI models, basically to improve 537 00:32:29,640 --> 00:32:33,080 their search capabilities to build smarter recommendation 538 00:32:33,080 --> 00:32:37,080 systems and even power enterprise chat bots with more 539 00:32:37,080 --> 00:32:41,360 reliable information. For example, Again, stepping 540 00:32:41,360 --> 00:32:45,520 back to the retailer shoes here, if a retailer is using HANA 541 00:32:45,520 --> 00:32:50,360 Cloud, he could now connect customer purchase history, 542 00:32:50,880 --> 00:32:55,840 product categories and user behaviors to build an AI driven 543 00:32:55,840 --> 00:32:59,240 recommendation engine that would understand the relationship 544 00:32:59,440 --> 00:33:03,880 rather than just relying on basic plain filtering kind of 545 00:33:04,240 --> 00:33:08,000 operations. So in short, yes, HANA Cloud's 546 00:33:08,000 --> 00:33:12,480 Knowledge Graph engine can be used exactly in these ways, like 547 00:33:12,480 --> 00:33:16,240 making the systems much more faster, more accurate, and 548 00:33:16,240 --> 00:33:19,840 providing them, providing them with more and more context so 549 00:33:19,840 --> 00:33:21,720 that they do not hallucinate at all. 550 00:33:23,440 --> 00:33:26,600 That sounds great. So thanks for introducing and 551 00:33:26,600 --> 00:33:29,080 explaining all that. When we look a little bit in the 552 00:33:29,080 --> 00:33:32,960 future, I know it's hard, but when we look into the future, So 553 00:33:32,960 --> 00:33:36,520 what do you see? How knowledge grabs are evolving 554 00:33:36,720 --> 00:33:39,400 and what exciting developments should we look out for? 555 00:33:40,720 --> 00:33:44,040 That's such an exciting question looking ahead in the future 556 00:33:44,040 --> 00:33:48,400 because knowledge crafts are evolving rapidly and we are only 557 00:33:48,600 --> 00:33:51,320 scratching the surface here of what they can do. 558 00:33:51,320 --> 00:33:55,240 We are just beginning. I would say right now, knowledge 559 00:33:55,240 --> 00:34:00,480 crafts are already helping to transform search, AI 560 00:34:00,680 --> 00:34:03,640 recommendation systems and fraud detection systems. 561 00:34:04,360 --> 00:34:08,480 But in future, I see them becoming even more tightly 562 00:34:08,480 --> 00:34:12,520 integrated with the real time AI or the automation systems and 563 00:34:12,600 --> 00:34:14,719 autonomous decision making systems. 564 00:34:15,520 --> 00:34:20,520 One major trend that I foresee is the fusion of knowledge 565 00:34:20,520 --> 00:34:23,239 graphs with the large language models, which is already 566 00:34:23,239 --> 00:34:26,560 starting to happen. I would say so the LLMS are 567 00:34:26,560 --> 00:34:29,159 really great at generating text, right? 568 00:34:29,159 --> 00:34:32,679 Like I mentioned, they sometimes do struggle with factual 569 00:34:32,679 --> 00:34:37,760 accuracy, but when they're paid with knowledge graphs, AIS can 570 00:34:37,880 --> 00:34:42,880 pull off a lot of things from a structured, verified knowledge 571 00:34:42,880 --> 00:34:47,480 base and then they can even make responses more precise and 572 00:34:47,480 --> 00:34:49,719 reliable. So this will be huge for 573 00:34:49,719 --> 00:34:54,520 applications like enterprise AI, chat bots, virtual assistants, 574 00:34:54,520 --> 00:34:57,520 and even intelligent decision making systems. 575 00:34:58,160 --> 00:35:02,680 And another exciting development that I foresee is graph based AI 576 00:35:02,680 --> 00:35:05,600 reasoning. Like instead of just analyzing 577 00:35:05,600 --> 00:35:10,040 historical data, the future AI models will use knowledge graphs 578 00:35:10,280 --> 00:35:15,280 to predict the outcomes and suggest next best actions 579 00:35:15,680 --> 00:35:19,360 whether in finance, healthcare or supply chain management. 580 00:35:19,360 --> 00:35:23,880 Again, just as an example, so you can imagine an AI system 581 00:35:23,880 --> 00:35:27,680 that is not only detecting the fraud, but also is predicting 582 00:35:27,680 --> 00:35:31,400 which suppliers might be at risk of disruption because of all 583 00:35:31,400 --> 00:35:34,720 this based on the connected data that you will have in hand. 584 00:35:35,520 --> 00:35:38,840 And then there is real time streaming knowledge graphs. 585 00:35:39,080 --> 00:35:42,640 So today many graphs are built from static data, right? 586 00:35:42,640 --> 00:35:46,080 But we are going to be more dynamic in the future, like in 587 00:35:46,080 --> 00:35:51,160 the sense real time graphs that needs continuous updates because 588 00:35:51,160 --> 00:35:53,520 the new data is going to flow in all the time. 589 00:35:53,720 --> 00:35:56,440 So this will definitely power the real time recommendation 590 00:35:56,440 --> 00:36:01,400 systems, fraud detection as well and also adaptive AI systems to 591 00:36:01,400 --> 00:36:03,720 help them make the decisions on the fly. 592 00:36:04,480 --> 00:36:08,520 And just tying it back to SAP HANA Cloud, I think the future 593 00:36:08,520 --> 00:36:12,560 is incredibly exciting because with the knowledge graph engine 594 00:36:12,720 --> 00:36:16,520 now being the part of the multi model family, businesses can 595 00:36:16,520 --> 00:36:19,240 definitely combine their structured, unstructured and 596 00:36:19,240 --> 00:36:23,280 semi structured data and also graph data in base that were not 597 00:36:23,280 --> 00:36:27,640 possible before. So as AI adoption is growing, I 598 00:36:27,640 --> 00:36:31,800 see knowledge graph engine in HANA Cloud becoming a key tool 599 00:36:31,800 --> 00:36:35,880 for covering these matter applications that the industries 600 00:36:35,880 --> 00:36:39,480 are planning to build. Just to sum it up, the knowledge 601 00:36:39,480 --> 00:36:43,360 graphs are not just about connecting data anymore, they 602 00:36:43,360 --> 00:36:46,240 are really becoming the foundation for next generation 603 00:36:46,240 --> 00:36:50,680 of AI driven intelligence. All right. 604 00:36:50,920 --> 00:36:53,360 Thank you for looking into the future for us. 605 00:36:53,640 --> 00:36:57,720 The key sense of today's episode is on the one hand side to learn 606 00:36:57,720 --> 00:37:00,760 what a knowledge graph is, but on the other hand side, of 607 00:37:00,760 --> 00:37:04,480 course the great news that you can now make use of knowledge 608 00:37:04,480 --> 00:37:08,280 graphs inside of SAP HANA Cloud. So if someone is interested in 609 00:37:08,280 --> 00:37:12,640 learning more about knowledge graphs within SAP HANA Cloud, 610 00:37:12,800 --> 00:37:16,760 what are some first good steps, tools, or resources all 611 00:37:16,760 --> 00:37:20,320 listeners could check out? Yeah, absolutely. 612 00:37:20,320 --> 00:37:23,600 If you are interested in exploring the Knowledge Graph 613 00:37:23,600 --> 00:37:27,240 engine in HANA Cloud, there are plenty of resources to help you 614 00:37:27,240 --> 00:37:29,680 get started. First and foremost, I would 615 00:37:29,680 --> 00:37:34,000 recommend checking out the SAP Help Portal page and the SAP 616 00:37:34,000 --> 00:37:37,880 HANA Cloud Documentation page. They have detailed guides on 617 00:37:37,880 --> 00:37:40,480 setting up and using the Knowledge Graph engine. 618 00:37:40,800 --> 00:37:44,920 And if you prefer a hands on approach, I believe the SAP HANA 619 00:37:44,920 --> 00:37:48,520 Cloud Basic Trial is a great way to experiment with the Knowledge 620 00:37:48,520 --> 00:37:52,280 graph queries and see how they will eventually work in real 621 00:37:52,280 --> 00:37:55,960 time scenarios. And for more interactive 622 00:37:55,960 --> 00:38:02,520 learning, SAPSAP is going to offer tutorials on SAP Discovery 623 00:38:02,520 --> 00:38:07,800 Center and developer focused contents on SAP Community where 624 00:38:07,800 --> 00:38:11,280 you can like maybe find blogs, discussions, and even some code 625 00:38:11,280 --> 00:38:14,280 samples. And if you want to do some deep 626 00:38:14,280 --> 00:38:18,280 dives, keep an eye out for SAP tech sessions, webinars, and 627 00:38:18,280 --> 00:38:21,920 other online courses because they definitely cover advanced 628 00:38:21,920 --> 00:38:27,400 use cases and best practices. So whether you prefer reading or 629 00:38:27,400 --> 00:38:31,480 you want to get some hands on learning or some community 630 00:38:31,480 --> 00:38:34,560 engagement, there's always something for everyone to get 631 00:38:34,560 --> 00:38:38,120 started. Exactly. 632 00:38:38,120 --> 00:38:41,320 So thanks for sharing. And it's only a call to action 633 00:38:41,320 --> 00:38:45,960 for every of our listeners to make yourself familiar with the 634 00:38:45,960 --> 00:38:49,320 new knowledge prep engine on handed cloud for some of the 635 00:38:49,320 --> 00:38:51,960 resources, also LinkedIn the show notes. 636 00:38:52,360 --> 00:38:54,920 And with that, we made the official part already. 637 00:38:54,920 --> 00:38:58,240 So thank you, Shabano, for having this conversation with me 638 00:38:58,240 --> 00:39:01,680 and sharing all details about knowledge crafts and Hannah 639 00:39:01,680 --> 00:39:04,560 Cloud. So this was a really good 640 00:39:04,560 --> 00:39:09,760 overview, I believe. And if you want, let's do 641 00:39:09,760 --> 00:39:12,960 something fun coming to the end of the podcast. 642 00:39:14,920 --> 00:39:18,800 So aside from our interview topic, I'm deeply passionate 643 00:39:18,800 --> 00:39:21,000 about listening and creating music. 644 00:39:21,320 --> 00:39:25,240 So therefore I play a music game in my podcast episode. 645 00:39:25,280 --> 00:39:29,800 So the game goes like this. I read an extract of lyrics from 646 00:39:29,800 --> 00:39:34,840 a famous song of a musician. I give you 3 artist names and 647 00:39:34,840 --> 00:39:38,560 you need to guess from which artist the lyrics we're from. 648 00:39:38,640 --> 00:39:41,800 And if you want, you can also name the song title of course. 649 00:39:42,240 --> 00:39:46,240 And of course, it's based on your taste of music you shared 650 00:39:46,240 --> 00:39:50,080 up front with me. If you need some help to finding 651 00:39:50,080 --> 00:39:53,200 the right answer, I can also give you some hints to the 652 00:39:53,200 --> 00:39:55,280 corresponding artists. That would be. 653 00:39:55,560 --> 00:39:56,720 Easy. Do you want to play the game 654 00:39:56,840 --> 00:39:58,320 with me? Let's let's get started. 655 00:39:58,680 --> 00:40:01,680 Let's pick on my memory. Let's see how that is. 656 00:40:02,000 --> 00:40:08,040 OK, so then the lyrics extract is I used to roll the dice, feel 657 00:40:08,040 --> 00:40:13,480 the fear in my enemy's eyes and the interruptions are A. 658 00:40:13,880 --> 00:40:22,080 No patrol B the weekend or C Coldplay and I said I can give 659 00:40:22,080 --> 00:40:26,000 you hints if you want. Oh this is definitely not 660 00:40:26,000 --> 00:40:31,640 weekend. I guess this could be Coldplay. 661 00:40:32,480 --> 00:40:35,080 Just guessing. Exactly, it's Coldplay, but did 662 00:40:35,080 --> 00:40:37,800 you also know the track name? I mean, I was. 663 00:40:37,800 --> 00:40:39,920 Not I. Was not asking for that. 664 00:40:40,080 --> 00:40:44,600 But I guessed it, so I don't definitely remember the title of 665 00:40:44,600 --> 00:40:45,360 the song. Yeah. 666 00:40:46,160 --> 00:40:49,360 So then maybe with one of my hints I've prepared, so 667 00:40:49,360 --> 00:40:53,920 translated the song title where the kroto the extract is out of 668 00:40:54,400 --> 00:41:00,240 the song title would mean Live the life, maybe now. 669 00:41:00,240 --> 00:41:04,960 Oh, OK, yeah, I mean, it's Viva La Vida. 670 00:41:04,960 --> 00:41:06,320 Viva. Yeah, Viva La Vida. 671 00:41:07,160 --> 00:41:09,720 Exactly. I have prepared two other hints 672 00:41:09,720 --> 00:41:13,640 just to make it complete. Originally, Coldplay was founded 673 00:41:13,640 --> 00:41:17,920 under the name Starfish. I did not know that before. 674 00:41:18,120 --> 00:41:22,400 And the band, founded in 1996, bought the real commercial 675 00:41:22,400 --> 00:41:27,480 breakthroughs beginning 2000. Yeah, thanks for playing the 676 00:41:27,480 --> 00:41:30,560 game. And we always ask our podcast 677 00:41:30,560 --> 00:41:33,000 guests for some recommendations at the end. 678 00:41:33,000 --> 00:41:36,680 So can you recommend a tech related book series or movie to 679 00:41:36,680 --> 00:41:38,520 our business? And if you have a 680 00:41:38,520 --> 00:41:41,880 recommendation, why do you think it's so good and nice to share? 681 00:41:43,360 --> 00:41:45,800 OK, I think we come to the end of the show. 682 00:41:45,880 --> 00:41:48,640 Thanks so much, Nicholas. I think it's so much fun. 683 00:41:48,640 --> 00:41:52,280 I had absolute pleasure talking everything about knowledge 684 00:41:52,280 --> 00:41:55,160 graphs and the knowledge graph engine and Hannah Cloud, and I 685 00:41:55,160 --> 00:41:57,040 believe this was a great conversation. 686 00:41:57,400 --> 00:42:01,360 Coming to the recommendation part, I think for a book I would 687 00:42:01,360 --> 00:42:05,080 recommend Life 3 dot O by Max Tegmark. 688 00:42:05,600 --> 00:42:07,720 I recently went through that book. 689 00:42:07,720 --> 00:42:11,720 So I believe it's just not for AI researchers. 690 00:42:12,160 --> 00:42:16,520 It's a fascinating look at how AI is shaping the future of work 691 00:42:16,520 --> 00:42:20,000 and the society and even human identity. 692 00:42:20,480 --> 00:42:24,920 So whether you are into AI or data science or just curious 693 00:42:24,920 --> 00:42:29,480 about where this technology is heading, this book would offer a 694 00:42:29,480 --> 00:42:34,160 thought provoking perspective on what an AI driven world would 695 00:42:34,160 --> 00:42:37,840 look like in the future. And I definitely want to add a 696 00:42:37,840 --> 00:42:40,640 movie tuck in here. As a recommendation. 697 00:42:40,880 --> 00:42:43,880 I would go with my all time favorite movie, The Imitation 698 00:42:43,880 --> 00:42:46,320 Game. I believe it's a brilliant story 699 00:42:46,320 --> 00:42:51,760 of Alan Turing and his work in code breaking and which kind of 700 00:42:51,760 --> 00:42:54,520 laid in foundation for modern computing. 701 00:42:54,520 --> 00:42:59,920 And AII believe it's a reminder all the time of how powerful 702 00:42:59,920 --> 00:43:04,800 problem solving and innovation can be and how you should be of 703 00:43:04,800 --> 00:43:07,680 the type of never give up kind of an attitude. 704 00:43:08,520 --> 00:43:10,800 So that's why I take away from the movie, and I think it would 705 00:43:10,800 --> 00:43:13,480 be a brilliant watch for anyone who's not watched it. 706 00:43:14,000 --> 00:43:15,960 Yeah, I think that's all from my side. 707 00:43:17,440 --> 00:43:19,680 Yeah, thank you, Shabana, also for sharing your 708 00:43:19,680 --> 00:43:22,480 recommendations. I definitely need to check out 709 00:43:22,480 --> 00:43:24,160 the movie because I never heard about it. 710 00:43:24,520 --> 00:43:27,640 And yeah, with this, we're at the end of the third episode of 711 00:43:27,640 --> 00:43:32,840 the year 2025, episode number 117 of our SAPPTP talk. 712 00:43:32,840 --> 00:43:37,840 And thanks again, big thanks to you, Shabana Sam Sudin, for 713 00:43:37,960 --> 00:43:41,440 joining me in the podcast today. I really appreciate that you're 714 00:43:41,440 --> 00:43:45,280 taking the time to talk with me. To our listeners, check out all 715 00:43:45,280 --> 00:43:47,760 the links in the show notes. And of course, make sure to 716 00:43:47,760 --> 00:43:51,480 follow Shabana on LinkedIn in case of any questions or if you 717 00:43:51,480 --> 00:43:54,320 want to share some feedback, drop us an e-mail at 718 00:43:54,320 --> 00:43:59,320 platformtalks@sap.com or just send us a message on social 719 00:43:59,320 --> 00:44:01,880 media. If you enjoy your podcast, of 720 00:44:01,880 --> 00:44:03,960 course, make sure to spread the word and don't forget to 721 00:44:03,960 --> 00:44:07,680 subscribe so that you don't miss one of the next episodes in the 722 00:44:07,680 --> 00:44:09,600 future. And with that, my name is 723 00:44:09,600 --> 00:44:12,360 Nicholas and I'm happy that I were your host for today's 724 00:44:12,360 --> 00:44:15,280 episode, recording from the SAP headquarters in the world of 725 00:44:15,280 --> 00:44:18,600 Germany and talking to you in one of the next episodes. 726 00:44:18,800 --> 00:44:19,320 Until then.