LIVE: Executive Keynote with Juergen Mueller | SAP TechEd in 2023 | Join Us!
Humanity has achieved remarkable Feats we've harnessed the power of Science and Technology to overcome countless challenges and connect with one another across the globe now look where we are today please welcome on stage yorgen mu Chief technology officer member of executive board sap welcome everyone to sap Tech 2023 welcome to Bangalore Bangalore and a very warm welcome to every single one of you joining us around the globe that was a very nice opening video right and that voice we heard was completely generated by AI since last sap teet we have seen a huge leap forward with the advancements of generative Ai and I think we are now starting to understand how generative AI will be Transforming Our World and I'm very sure that all of us have now experienced the magic of generative AI in one way or the other in our private lives I for example I recently received the uncle of the Year award and that is also thanks to Jen I because while I think my storytelling skills were already quite good my nieces have turned into even bigger fans now that and I now I'm getting a little bit of help from CET gbt but here at sap techet of course we look at how to apply this magic to business AI because the game is on for all of us all developers now also have to become AI developers and today we will show you what you can expect from sap and why the sap business technology platform is the best choice for you we will cover cover all pillars of btp we will look at application development we will look at automation integration data management analytics and Ai and by infusing generative AI into our Solutions and by giving all of you access to the technology and the learnings we have you will see that sa is your best friend to deploy relevant reliable and responsible generative AI in production and together with our partners the sap ecosystem will be the number one business AI ecosystem and I'm excited to have a wonderful lineup here today of customers partners and colleagues joining me on stage to bring our story to life I hope this will inspire you and that you will enjoy the show let me give you some sense of of scale and momentum of our business technology platform at sap a five-digit number of people is waking up every morning to making btp even better since last tet we have delivered more than 1,000 road map items year-to date availability is above 99.99% and last year I shared with you that we had 19,000 customers on bdp live in the cloud since then we've actually added 3,000 so we are now at more than 22,000 live btp customers in the cloud and what's even more important for a platform is adoption and that grew more than 60% year over year and with so many new customers and developers coming on of course we want to make the onboarding experience as quick and as smooth as possible so let me give you a sneak peek into how we are simplifying the btp admin and development experience with generative AI you will be able to ask our new co-pilot Jewel which is integrated right into your btp cockpit for example you can ask a question related to btp services and get up to-date recommendations you might then ask for help with onboarding and based on on your prompt a service entitlement is added to your btp account and as duwel is integrated across sap Solutions you can switch to other Solutions like sap start for example and continue to interact with btp I can for example onboard my colleague Michael Ling to the btp account right now from sap start and that context of your conversation is never lost As you move from one sip solution to the other using Jewel really for the first time ever you will be able to interact with the btp cockpit from anywhere you are without leaving the context of your sap Solutions and now actually you see that Michael is successfully added to the btp account with that correct role we'll also be able to create new accounts services sub descriptions and a lot more isn't that amazing and while we are working on making this available to you in the first half of next year you can already use our terraform provider today to automate the setup of your btp environment and as the terraform provider is based on the btp command line interface it covers actually all its apis managing director softw accounts subscriptions etc etc and now before we dig deeper into the btp pillars I showed I want to invite two guests on stage and let's start with the first one KK from HCL Tech k k very very good having you nice to see you man great so every customer and partner today actually will get get more or less the same question and that is why is sap btp your choice and what is next for HCL Tech KK thanks huran for having me and good morning good good morning and people who are watching across other parts good afternoon or good evening uh thanks for having me here and the first thing on why we doing the sap btp I think the traditionally people have tried to mess around too much with the core of many of this critical Enterprise systems and value innovation has never happened by trying to disrupt the core and the value which we're seeing with btp is that you can continue to innovate and expend ex extend we've done some amazing work and now both across core data engineering and now adding lot of generative AI capability which is available on top of the sap this thing it's been an amazing collaboration between the sap team with our uh Solutions team within HCL and also a lot of our other aspects of our business so as a sap strategic Global partner as a GSI but also with software which is our isv uh business where we are building lot of software products on top of the sap btp platform and extending lot of our product lines one of the interesting examples which we had was the automated sales return capability a lot of times we try to use tools but where we really see is that you have to bake generative AI as part of the process or the function rather than trying to just give generative AI as just I want to use it it's it's about how do you really Drive business outcome and benefits around that and we've seen that in this use case when you have a lot of manual processes or steps and most of the times people think humans operate on workflows we love interactions we love conversations that's how we really engage and with the power of generative AI you can really change the way EXP experiences across front office middle office back office the whole integration layer and that's that's the big difference you can actually humanize technology by really bringing conversations and Bing them into the way you do your uh stuff the second interesting uh piece which we've been trying to uh build is that this whole concept of value extraction from existing processes and systems rather than what we see today a lot of mindless app modernization people people are modernizing apps without knowing what's the end destination I think it's very very important to see if you have invested in something extract the value of what you have and then build new application so either you build or you extract value modernize only if there is really a need to modernize and I think that sense I think the developer Community will start to realize that keeping on tinkering Old systems versus building new another interesting example this is something very close to uh our hearts is what we call osphere this is a solution which was built on btp but the inspiration for this solution really came in from our chairperson's big initiative called the aqua pruner which was launched with the world economic Forum up link this is about how do you really use intelligence to understand the usage of water the IR is that 70% of the earth is surrounded by water we are 70% water water but we still have water shortage so there's abundance on one end but again data metrics and ability to bring those pieces together is going to be extremely important and with the aquasphere the solution what we really have built is use intelligent Hardware which can be at the source at the consumption layer but all this data gets merged with S4 Hana and then extend a lot of those intelligence capability using btp we've now also infused some level of gen capability so in this example what we did was we ingested our own sustainability report and said okay get me context I think the most important thing which we realize in gen is to build context and that context building allows you to really ask the right question or even ask close enough a right question be able to extract responses and make it more uh usable for the users so by combining lot of these pieces we are able to really deliver an outcome to to a customer and as the llm improves correct as you really start to uh take this whole piece out you'll start to realize that more and more we apply large language models and then mix it up with metrics data your ability to really extract business value and outcome is going to be extremely important K K thanks for sharing then thank you Oran for having me that's good okay thanks a lot bye and talking about um water scarcity actually the next um person and guest I will ask on stage here gave me permission to share that she actually faced challenges accessing clean water when growing up in Brazil but let's hear from her directly please welcome Thea from the social startup sustainable development and waterfall sdw hi you thanks for coming all the way from Brazil and sharing your story with us same question to you why is btp the choice for you and what is next for sdw oh I'm so happy to be here thank you so I'm from small toall in Brazil where access to water and sanitation is difficult I grew up seeing people not having a sewage system and often not even having access to a bathroom tap water was also not a reality in many cases we uh water being stored and containers and tanks should guarantee a certain volume of water and most of the time this water was not purified becoming a potential source of disease changing this situation has always been a dream that has become real with as the double we work to provide service and device for communities that live without access to Safe water and sanitation our main products uses solar energy to purify water and until now s dou brings clean water to 23,000 people and we help it to save more than 6 million lers of water with our sanitation products being a startup we must put a lot of effort to get funding and continue our social calls the customers need to see a return on investment so they want to know how their money is spent how many communities and people was served and how many products were installed to avoid waterbor disease so we need to prove that what to do with our customers money has the highest impact we have developed an application using sap build apps it is used by employers to interact with local communities and to enter data about their sanitation services and social parameters the application is been used to Missouri kpis and the social impact of the services our F workers provide by using the app we can provide our clients with upto-date data at their fingertips in addition we can ensure data safety and confidentiality thanks to the app built on sap btp yeahia thanks for sharing your story thank you again and it's amazing what you do with sdw thank you thank you so um now let us continue actually with the first pillar of the business technology platform and um I want to hand over this stage actually to our Michael aming to talk about application development Michael St thanks a lotan and good day everyone so we have just seen how easy it is to onward onto btp and set up your btp landscape the next logical step is to use it for developing extensions and applications why is btp the choice for just that developing side-by-side extensions on btp helps you to keep the core clean this way your systems like an sap S4 hun are always easy to upgrade last year we ounc sap build our loow code solution for development and automation it has become one of the fastest growing btp Solutions and it's also part of rise and grow with sap Andy just heard from sdw that sap btp helps to make a difference in the world already by providing access to water and now we expanding our set of professional development application development tools on sap btp today I'm excited to announce our new procode offering sap build code sap build and sap build code provide you with a simplified developer experience a unified Lobby a starting point for sap build and sap build code tools for enhanced collaboration between citizen developers and professional developers often referred as Fusion development and new generative AI capabilities to increase developer productivity of course sap build follows a standard programming model for Java and JavaScript based development it uses a cloud application programming model in short cup but enough talking let's see it in action now and for that I'd like to welcome Sangita on stage Sangita will create an app live on stage and this app will allow you to register for the preview version of sap build code Sangita the stage is yours thanks Michael we are now in the unified Lobby of sap build and sap build code what Michael just talked about it is the central entry point for citizen and professional developers to create applications automate processes and build business sites I use it today as well to create a registration app live for you to get early access to the sap build code preview if I click on the full stack application you get the services and data annotation ready for you to save us some time I have already pre-created an empty project we are now in the sap build code environment which is based on sa business application Studio our cloud-based IDE to develop extensions and Solutions on sap btp the new and cool thing is that we now have Jewel our AI co-pilot fully integrated for you to give a kickart with your development today I'm going to ask jel to create an app for the registration for you today jul is now thinking this looks pretty simple San and while AI is preparing for this app let me highlight a few points jewel is fully integrated into sap start in the btp cockpit and in sap build code and over the next two days you will see many examples how we integrated Jew across the entire application portfolio Jew uses generative AI capabilities as well as Miss machine learning and natural language processing it quickly sorts through and puts data from multiple sources into context for example developer documentation code samples and the like and finally it surfaces the insights to help us as coders to do our work faster and in generated code has built in cap best practices Sita is Je already ready yes it is you can see jeel has already already generated the code for me I have accepted it and you can see the data models services and the UI annotations ready for you let me quickly switch to the graphical modeler and now I want to add additional business logic so I want to restrict the number of users who can register today for the sap Bill code Pro preview I opened the code Editor to to add the business logic I also Define the events when the business logic has to be executed and here as well I again Ask to so I ask Jewel to limit the number of users until the limit is reached jewel is allowing you to register after that it says Hey the list limit is done jwel is again thinking oh tool has already generated the logic for me I can just click and copy paste it so now let me also get a UI application on top for us to preview how the app would look like I go through the visard and create a template based browser based application and then click finish I can easily click on run and quickly preview our application yeah and while we are at the final step for creating this app let me explain actually what you did not see on the screen in the back sap build code makes use of cap for the creation of this app and this has a number of benefits the CDs models are created automatically and the standard interfaces get cre created as well and by the way an OD data API was created during this process so once deployed the apis can be consumed by other tools like sap build low code Solutions do we see the preview already yeah Michael be a little patient uh julus just generated the code we used it and the app is getting built you can see that it's being built give it a bit of a t I'm patient yeah cool and you see uh that we have know the preview option here of course now we could deploy the app but let's open a predeployment now I want invite everyone to register for the prev version of sap build code with the app we have just created yes this is real just scan the QR code and we will reach out to you thank you Sangita so much give an Applause with sap build code it took us just 5 minutes to create a full-blown applications with quad operations and high quality code based on cut that's pretty amazing but let me stress again with sap build code you can maximize the value of your data and offer secure access topics like identity manag and audit logging and much more are automatically taken care of sap build code is already available as a preview offering and you can subscribe via the QR code as you have seen us here now sap built sa buil code and sap btp abup environment can be now used together for low code and proc code development talking about abup let's now look at some new features for ABA Poco Developers last year we announced ABAB Cloud for proc code development on btp and S4 today we are adding AI to it let me show you a preview for delivering AI assisted ABAB Cloud programming in this example I want to create an app that helps me to keep track of developer conferences we see how AI assisted code generation is also built to up up development M tools in the generator you simply enter user prompt explaining the app and we see the artifacts are being generated and of course this is all written to AB transports and here we see already some sample data and of course we find our session sap teet Bangalore now another great thing about abup Cloud you can use it for writing code in btp and inside us for Hana system following the clean Core principles now this was just a short tool demo but just imagine how powerful this is will be for your own business use case for example how AI helps you to explain and review upup code generate test cases test data and much more now I have one more thing for you having seen this how many of you want to know about the best way develop extensions applications on btp yeah I see a lot of people are curious great I have one more thing for you from now on there's one Central source for best practices for your procode development Journey the sap btp developers guide this guide explains you how to build extensions and apps with sap build code and sap btp upup environment the sap btp developers guide is available for everyone since today on help.sap.com so check it out scan the QR code and with that I'm heading back to Jurgen thank you everyone thank you Michael so as you can see we have a lot of highlights for you in that area of application development let me just quickly summarize sap build code provides you with a simplified developer experience through a unified Lobby and this makes collaboration between citizen developers and professional developers very easy and we are infusing generative AI capabilities to increase developer productivity we also help you to follow the latest best practices with our sap btp developers guide and sap build in the latest addition to the build family sap build code and the sap btp upup environment can now be used together to create apps and extensions and doing that is of course extremely important in the context of a clean Core strategy with Espana you develop extensions either on btp or in Espana with ABA cloud and while a clean Core is not a prerequisite for esana it makes any esana upgrade of course so much easier and let's hear from a customer who has lots of experience in applying the clean Core strategy please welcome Hitachi Hightech Sakai son [Music] Sak great having you I'm very excited to with you in bangalo thank you very much same question to you yes why is btp your choice and what is next for Hitachi Hightech yes okay uh let me explain about my journey our company has implemented sap Erp globally long time but over the last 20 years system become complex and as many as 9,000 areon for individual optimization for each region this complex system required significant maintenance cost and prevented company from quickly introducing new technologies and contributing the company's growth so it is very critical that we reduce the complexity uh thanks to btp sap btp side by side extension we can now easily upgrade our s so sap S400 systems by year or by turn so we can keep S400 cor clean and we can get the benefit from all latest sap Innovations we create over 800 new addons for UI and report and interface as a fact you know we used to have 9,000 adoms and we reduce to 8 100 addons in addition we are using sap btp to automate internal Global end to end process and we are planning to expand our previous experience with robotics process Automation and aim at automating data entry using sap built process automation solution vro management and RPA capability en laed cross system workf fors in addition sap build have been initiated e efficient development and also we are planning to enhance process integration between sap S400 Cloud public and private using sap event mesh of course in the future we are also planned to ji to data driv management uh we are using sap btp I think using sap btp we can build Global digital process smart Global digital process uh without touching sap S400 core so that's is very very incredible thing we believe sap btp could be very very important tool to accelerate our business process thank you thank you so much for sharing your story thanks a lot thank you you just heard from sakaan about the clean Core strategy but also how important Automation and integration are so let's talk about Automation and integration and let's face it everything that can and should be automated will also be automated especially now in the generative AI age and we help with sap signavio so you can constantly check your process sees and identify opportunities for automation with our latest release of signavio actually you can use btp to connect to sap or thirdparty systems for process Mining and process mining events from signavio can now be published to sap integration Suite event mesh sakaan just talked about it this allows you to immediately react to process anomalies as they happen and the best way to react to process anomalies and to automate processes is with sap buil and the combination of sap signavio of sap build integration of sap build and of the of the sap integration Suite allow for holistic Enterprise automation an Enterprise automation helps you to identify to model and automate business and it processes with our AI powered tools and with our recently announced intent to acquire lean ex they are a leader in Enterprise architecture management software we are expanding our business transformation portfolio even further and now let's hear from one of our longterm customers about their Automation and integration journey and for that I would like to welcome rucha on stage from the Mahindra group gooda it's great having you it is great having being here you know our our journey has been 25 years with sap so thrilled to be here thank you it's very very true Roa same question to you why is btp the choice and what's next for the Mahindra group um allow me for a second to talk about Mahindra group a little bit um we are a Purpose Driven company we are a a global company um our our purpose is to drive positive change in the communities we are in in the lives of people who live in the communities we believe that we will only rise if we enable others to rise and that's our philosophy uh you know I I think rise resonates with rise uh the three pillars of our philosophy I want to talk for a second about one is rise for a more equal world second Rise to be Future Ready third is rise to create value now it is great to be speaking about technological innovation as you can see it fits squarely in our Central pillar rise to be future ready uh talking about btp Jurgen we started identifying tasks that don't require a lot of human intelligence and that enabled our human forces to work on more strategic initiatives we have used used btp build process automation to automate more than 250 business processes in 3 months that's the agility that has come through it uh major work has been done actually the the great part of this is most of this has been done by our citizen developers we've trained 100 plus citizen developers and they have helped us do this um integration Suite has helped us connect uh our entire ecosystem sap nonsp CS third party and our homegrown applications as well um it has helped reimagine the Integrations move away from file based transfers even apis and go to realtime connectivity data exchange through event mesh uh integration Suite API uh management has helped reusability um we embarked on a major modernization and transformation Journey about a year ago and sap has been a big part of our journey and reusability was a major measure M for us to be able to standardize and so on uh API management has helped a great deal we have about 100 apis on boarded um with all the policies around caching rate limiting authorization Etc let's talk about the outcome this is all great the work goes on what's the outcome 900 Plus work days saved 400 plus developer days development days saved in a year's time I think we already talked talked about citizen developers and I think that Community is growing by the day um some more numbers 15% Improvement in high compute transaction uh some of you will resonate with this right our month and our year and we want to do it as fast as possible we also saw 35% gain in our developer efficiency we have seen 20% faster um execution of our jobs and this is my favorite number we have saved over 13,000 trees in a year I think we all all care about leaving the planet in a better place uh this is this is a great achievement we also have been partnering uh with sabtv for a lot of AI use cases uh I think we're all in the generative AI bus but AI at large is a big part of our work um I'll give I'll talk about a couple of examples one is around AI based invoice scanning uh we had been doing this with OCR but we all know the challenges that come with it um we are moving it away from uh traditional ACR to AI based document uh to help with prediction and feature extraction of these invoices and it's going great um another one which is uh under consideration and very very important for us is use AI to understand customer repair requirements translate and propose possible rep repair Solutions um as part of a warranty job card analytics uh we're very excited about this one couple of more examples um identify legal risks in your documents in your contracts um use uh intelligent document classification uh your ID card comparison with what's in your success factor we are also success factor users and compare the employee records and maybe nudge if it is not the same to say let's have the uh latest data and so on so forth um I could go on and on about our our Ai and generative AI use cases but I think that should sum up why we use it BTV this is amazing rucha big thanks of thank you great to be here rucha thank you so so much great having you yeah we've just heard from rucher about generative Ai and in the context of Enterprise automation we talked about I want to show you what will be possible related to process generation and for that let's once more think about sdw the social startup we had earlier here on stage and let's assume they continue to scale and at a certain point in time they want to optimize their validation and approval process for projects so sdw could actually quickly ask our co-pilot Jewel in natural language to generate a process and you can get recommendations of pre-build content from the sap business accelerator Hub or you have the option to create a completely new process and that newly created process can be previewed and accepted and from that starting point actually you could enhance the process in sap build process automation you can modify the process for example the business rules and Define business process approvers using Jewel again now that process is updated and it's always now the right uh approver is selected depending on the category of that project and the location where such a project will be executed and with that sdw um would receive their Investments along more quickly and more efficiently so you saw how the use of generative AI will also accelerate process Automation and I know there are a lot of processes out there that need optimization we plan to make that available in the first half of 2024 I have one more important announcement in the area of integration quick show of hands here in the audience um which of your companies are using either sap integration Suite or sap process orchestration or sap process integration quick show of hands I thought that many many hands gr up um you will be very happy today because today we are elevating our sap integration Suite with the general availability of edge integration cell some of you already know what it is and you like it already let me explain for all the others Edge integration cell is our hybrid runtime so you can now design your processes in integration scenarios once in the cloud and then you can decide to run them anywhere you want and now even in private Landscapes such as remote factories or warehouses and this closes the last remaining critical gap for our process integration process orchestration on Prem customers to quickly move to our sap integration suite and with integration Suite you heard it today as well from rucha for example you can leverage more than 250 outof the Box connectors you can use event driven and API Le Integrations as well so let's summarize the highlights we had here in the area of Automation and integration with sap signavio sap integration Suite sap build they deliver a comprehensive Enterprise automation suite and we will embed generative AI into all of them and you don't need generic workflow Solutions like service now to automate processes in sap you also don't need generic integration tools like sales forces moft or Del boomi for the integration of sap with other systems so we just heard how important it is to simplify an automate business processes and what's related to every business process data so let's take a closer look at the topic of data management now however before I dive into our exciting updates let's hear from our from one of our esteem technology Partners the women's tennis Association actually let's draw the [Music] video hello sap dead 2023 sending warm wishes to you all from Florida to Bangalore the women's tennis Association was founded 50 years ago to run the business of women's tennis and serve players worldwide since 1973 the WTA has amassed Rich stores of data which were segmented across numerous databases to make the most of all of this data we needed to rearchitecturing sap btp is allowing us to translate Decades of data into holistic insights integrated statistics and Predictive Analytics on one unified platform this transformation project was also the Catalyst for the WTA to explore the world of AI working with sap we developed a prototype that utilized the AI capabilities of the wta's new btp landscape to analyze historical player statistics and find correlations that help predict future success the learnings we took from this project will now determine how we move forward player health and injury prevention enhanced tournament Logistics and leveraging dynamic data are just some of the topics we look forward to exploring as part of our data management strategy with sap we are so proud to not only Advance the sport of women's tennis through technology but also to Pioneer this new AI era alongside our longtime partner sap this is only the beginning and we are Beyond excited for what the future entails isn't that inspiring I didn't know that you can do that with tennis data I like the hidden Champion um prediction model the most we of course understand how important it is to collect and to utilize data efficiently and that is why we introduced sap data sphere in March this year a data service built on sap Hana Cloud sap data sphere creates a business data fabric to simplify data management of all your critical data for many companies probably also for your company maximizing the value of your data is very complex and time consuming why is that because too many companies got convinced that it somehow would make sense to export all sap data and then rebuild the context about your business processes again and again and again and and let me be very clear this is a misconception it only creates data inconsistencies huge extra effort a lot of frustration and cost but it doesn't have to be because sap data sphere unifies all your data into a single semantic model so in the end you have all your sap data exposed and governed via sap data sphere and then you connect to all your other data sources that you have and to accomplish this Vision we know that sap data needs to be open and work with your existing data infrastructure and this is why we announce strategic bidirectional integration with open data Partners such as kalibra confluent data breaks Google and data robot sap data sphere Innovations include a low code interface for data modeling a catalog with data semantics and lineage capabilities and Federation and replication capabilities for data integration since its launch we have also expanded our data center coverage on AWS on Azure and on gcp and today we announce sap data sphere semantic onboarding with sap data sphere semantic onboarding you can onboard sap application data and data products with full context and meaning to your business data fabric there's no need to recreate the semantics the logic the context that already exists for example in sahana or in sip success factors and this also allows business users to create data models without understanding the technical underlying dat data structures of our applications and this fresh new approach enables self-service and secure distribution of data within your organization and also outside of your organization with the sap data sphere Marketplace sap data sphere semantic layer allows you to easily start new analytics as well as planning projects and to quickly create intelligent data applications and that is the keyword for sap Hana Cloud sap Hana cloud is one of the most adopted Services internally at sap more than 165 different applications and services use sap Han cloud with its multimodel capabilities and now Hana Cloud takes center stage in the generative AI age I couldn't be more excited to announce sap Cloud's Vector engine this is huge and here is why we are working with all kinds of large language models like gp4 Lama 2 Falcon 40b claw 2 and many others for a while and they are amazing but they also have very sever limitations like outdated training data and business process context so let's make it very tangible imagine a large language model in llm would be a colleague of yours and that colleague would be very intelligent the colleague could program actually pass medical exams pass law exams can argue and has good manners it's a pretty good colleague huh but such a colleague would not know anything about what happened in the world in the last year and the colle would have no clue about your internal processes in your company nor any of your company systems like Financial systems HR systems and the like and I think this is the worst after every conversation you have with that colleague the colleague would forget what you talked about so working with such a colleague would be a little weird right and of like let's say limited value but try it out try asking a large language Model A seemingly easy question what do you think of the offer from our most important supplier last week of course an llm cannot answer such questions as they work on their initial training data all other data has to be provided as context and this is where saana Cloud's Vector engine steps in it helps to combine the powerful cap capabilities of large language models with all the relevant data of your organization this process is called retrieval augmented generation and sap is leading the way by bringing Vector capabilities into our industry leading sap Han Cloud database So within new sap H Cloud Vector engine you can really tap into all your mission critical data while keeping all of the amazing multimodel database capabilities of Hana Cloud but let's see it in action now and for that I'm happy to have Sabrina on stage thank you Jan so as you just heard sap Hana Cloud's new Vector engine will give you the ability to incorporate the data's underlying meaning into your intelligent data applications large large language models have been trained on the vast amount of data available on the internet like text or images and now have an implicit understanding of semantics and relationships so in the simplified graphic here we consider words collapsed in a simple two-dimensional plane you can see that the words database and table are closed together in our Vector space signifying that they are semantically similar while we have a much larger distance to the word Icebreaker sign in that it is not similar so here's how this works we live in a world that is three-dimensional but embedding vectors exist in a high-dimensional space with hundreds or thousands of Dimensions vectors are a mathematical data representation of any object in a multi-dimensional vector space they utilize values along a large number of Dimensions to describe the attributes or features of the object these Vector embeddings are generated by an embedding function that is specifically trained for certain object types in essence the so-called embedding functions convert complex and often unstructured data into a vector representation and preserve semantic similarity this enables you to process compare and use it in various machine learning and data analysis tasks using large language models so the great news for developers is that SQL knowledge is all you need let's see the vector engine in action here in the Hana database Explorer this table in the Hana database Explorer is leveraging the new data type real Vector we have been populating the table with the text of the sapana cloud graph engine documentation and to be honest nobody really likes to read extensive product documentations so what if you could just ask the large language model to summarize the documentation and extract specific information that is relevant to your question so we have used an text embedding function to generate these embedding vectors the embedding function Maps semantically similar text to vectors close to each other in a high dimensional space running a semantic surge then simply becomes a nearest neighbor surch in that Vector space with 1,536 Dimensions so where it differentiates is you're not searching content you're searching for condensed semantic meaning now we are running a nearest neighbor search in the Hana using SQL we're using the built-in distance metric cosine similarity that compares the vector in the table to the real query Vector the query Vector was obtained by the embedding function that is based on the string how how do I run the shortest path algorithm so now the result brings back text from our product documentation that is semantically similar to our query about shortest paths since this nearest neighbor search is a SQL operation we can add additional SQL operations for example we can set a filter now we're able to see the outcome on the bottom left and of course you can also do joints for example to the author masasa data table where the location of our authors is being stored so let's assume you would like to invite your experts that are located nearby to an in-person customer meeting we can then use a spatial predicate of the Hana spatial engine within distance to identify authors that are located in a two 200 kilm radius around Pune now taking a look at our location we can see on the map where these two dots are displayed our author's location is being stored so you must stay tuned now to hear more about this powerful Vector engine in combination with large language models later in this keynote thank you ggen thanks Saina thank you very much Saina I don't know if I going to have a headache tonight like thinking about a space that is more than 1,500 Dimensions but actually it's good that we have Hana Cloud for that so we are using the Hana Cloud Vector engine internally already today and as Sabrina said we couldn't be more excited to bringing that to you in q1 of 2024 so let's summarize the data management section here's what you need to take away first of all sap Hana Cloud's new Vector in engine allows you to achieve the ultimate goal in the context of generative AI combining powerful large language models with your data then sip data sphere delivers an integrated semantically Rich data layer across all your data data Landscapes no export of sap data is needed anymore and having all data accessible with sap data sphere this of course builds the foundation for talking about analytics so let's take a closer look and at analytics and who could do that better than our Global head of engineering for analytics she is based here in Bangalore let's welcome darani danani the stage is yours thank you so much Yogen for having me on stage it's my pleasure to talk about analytics Cloud sap analytics Cloud allows you to analyze your business data and understand what has happened it also allows you to simulate and predict future outcomes and it also allows you to create or adjust your entire organizational and department departmental plan for better future and the best part of it is that the entire analytics is being done on top of the live data coming in from SAP systems with that I would also like to say what's new first and foremost we did it again sap analytics Cloud's performance is better than before we are introducing lightweight viewer which provides up to 63% of faster rendering on top of the classical stack lightweight viewer is already in preview and it would be available as GA in q1 of 2024 and it also does not require any single migration steps and it provides a very seamless user experience to your existing stories by the way if you're still using classical stack dashboards get your dashboards move to the new stack to leverage all the performance improvements I talked about and also to leverage n number of optimizations and Innovations we've already built onto the new stack one of a customer a large multinational in the manufacturing industry who also had a huge sap footprint of saps for Hannah SAP BW for Hannah sap Hannah did a competitive study with sap analytics cloud and another leading bi vendor the study that was done with our sap partner interops found that even after simplifying the landscape and the scenario for the competitor sap analytics cloud outperformed in all the scenarios it provided up to 60% improvements in filtering 74 percentage improvements in the initial load and multi times on the navigations from what you see for doing analytics on live sap data there is no other better choice than sap analytics cloud now you saw how to get insights faster but what is really the quickest way to get your business questions answered is it clicking through screens no it's natural language of course we are super excited to introduce just ask into sap analytics Cloud which compliments Sap's AI co-pilot jeel with as you see here with just ask you can just simply type in your business questions and get your answers immediately just ask has been already tested by Beta by selected customers and it would be available for almost all of the Enterprise customers by q1 of 2024 going forward just ask would also be utilizing hanas Vector engine to Broad its knowledge Base by combining the vector data and Sap's relational data and that's not it we are continuing to invest on our new model that we delivered last year we are unifying the data first approach and the structured first approach to have a single unified enriched model and starting from q1 all of our planning and analytics users would be able to seamlessly use this new model in an endr W having said that let's also understand how we are planning to Leap Frog competition in the phace of analytics and more for that I would like to invite yogan back on [Music] stage thank you so much amazing thanks a lot thank you and dar you talked about leap frogging the competition actually we have an expert guest here the next one who really knows how to Leap Frog the competition with that let me welcome Michael from hankle on [Music] stage Michael welcome good to see you very good seeing you same question to you Michael why is BTV your choice and what's next for Han yeah thanks very much Jan also great to be here good morning to everyone happy to be here in Bangalore the Silicon Valley of India but coming back to your question jur sa btp why why is it our choice uh three years ago we were looking for large scale enterprise software technology platform which would help us to build sap applications side by side you guys call it or nonsp applications which would be out of the box integrated into the sap landscape and this is what we found with zap btp and even more Zap btp for the last three years has been our Innovation platform I have to say I'm very impressed what you guys have brought to the table in terms of geni super excited to hear all the announcement today and uh so we are going with z btp and it's really accelerating our Innovation pace so with that let me very briefly introduce hankle to you what are we doing and what's our digital Journey so hankle is a global company we have two major businesses one is a industrial business a B2B business model adhesive Technologies and we have a second business which is a consumer products Goods business so a classical b2c business and we as a company we have embarked about four years ago on a massive digital transformation journey and where we are leveraging software in particular in order to move the business so what we have been doing on the one side we have been building a huge internal software capability on the one hand we have by the way here in bangalow one of our biggest Innovation hubs across the world and on the second side we have been building strategic partnership and one of the most important ones is the one with sap uh we have been a long-term sap customer but we said for the next software era to come we need to much closer to partner with sap and accordingly we went into a partnership which we call digital Leap Frog and that's really uh has been paying off for us quite a bit so let me let me today talk about one specific uh case which is trade promotion management so trade promotion management for our industry is highly significant it's we are running a multibillion scope on on trade promotion management and it's really Mission critical for our business so three years ago we were looking into how to build it because we couldn't find anything in the market and we had failed with several approaches before so we said let's build it ourself together with SCP on the just newly released sub business technology platform it's a very very complicated application just to give you an idea not to go too deep you have to manage 2 billion planning nodes and which technology would be better suited to it also being fully integrated in the sap core then btp with Hana so that was working quite well we were able to deploy that across the entire organization at scale and now we are looking to the next uh innovation in this very very important area for us so before we go into that what we call Trade promotion optimization let me briefly talk about about how we think about J so gen for us is really super important we are believing that gen is a great opportunity for us uh I like to take here one quote which we always use this is Gen starts to disrupt economy and Society similar to the internet same magnitude like the internet but it's coming at light speed just look at the last 10 months since we had the inflection point last year in November what has happened just before sap has announced that they have now a vector storage which is able to manage large language model and it's really The Innovation pace is really mindblowing I have never been really discovering anything in my career so far an exponential technology which is kind of helping your business to improve exponentially I strongly believe that with Gen finally we have such a technology in place so coming now shifting over a little bit to the trade promotion use case again so with and you see a video in the background so trade promotion optimization is now all about the killer use case I mentioned to you before further optimizing that multi-million spend and so we were looking how to do that so we were building in an MVP tradition machine learning algorithms into the solution on the one hand and on the other hand we were trying out the just ask gen uh kind of human interface prompt so in order to make it available to our users so machine learning for optimizing which would include um the whole simulation planning uh part of it and then natural language processing in order to make it also available to the end user who is not so much familiar with very sophisticated blackbox Ai and that is really paying out I am super happy that sap and we were able to uh go into this Innovation Journey it was a 12 weeks MVP a lot of out of the box and just to give you perspective so we are not just using the uh newly released out of the box functionality but on the other hand btp is also super great for building our own gen Tech stack so with btp we have a technology in place by which we can build our own large language model llm and lmm IP which is super important for any company to build in your domain specific vertical knowledge and that was possible with in that project first of all I think because of the partnership secondly because we have a great technology in place which is innovative at a rapid pace so keeping up uh in high Pace then there was a great Innovation mindset very often we think big large tech companies like sap cannot be Innovative I have to tell you we saw exactly the different thing yeah and then last but not least because we have this trusted partnership we know not everything is working from day one so you can really jointly innovate um at Large Scale so with that I would like to also invite you to our boo we have a couple of colleagues here from sap but also from hankle also here from our Technology Center in Bangalore they have a boo where you can have a more detailed look into that and there is further great stuff available which we have build in MVPs uh with sap together with that thank you very much for your attention and I would like to hand it over to Yen again thank you thank you very very much Michael thank you so you've just learned about another important capability in analytics when you think about the last video you saw that is being able to do Advanced simulations and then adjust planning as well that is still in a private preview we are testing it with like 25 customers and partners in the moment so stay tuned you will hear more very very soon summarizing the analytics space we showed you two new important features features in sap analytics Cloud the lightweight viewer renders results faster than ever there's no better solution than sap analytics Cloud for sap data at least just ask is augmenting sap analytics cloud with natural language support and in total since last deade we delivered more than 150 features also we have more than 240 content packages available for sap analytics cloud and planning such that you can start your project right away with the best practices and of course we are integrating generative AI into the major workflows of sap and Antics Cloud you saw and heard a lot already about Ai and to round it up let's look at the remaining pillar of btp which is AI we want to show you how you as a developer can change the world with sap business Ai and for that I'm happy to give the stage to Walter Sun our head of AI hello morning welome thank you very much Jan for the warm welcome we appreciate it good morning Tech Bangalore great to be here yes thank you thank you I'm excited to share the new capabilities we're making available to developers as technology enthusiasts we're all aware that the hot summer of generative AI is extending to the Autumn and Beyond on artificial intelligence and particularly generative AI continues to reinvent how we run our businesses in many innovative ways that's why we believe this new technology will alter the way engineers and developers work in three ways one how we build into our software and applications two how how we took take what we build into the devop cycle tools and platform and three who does the building namely as natural language conversations extend programming Beyond low coders to even no coders development will become even more democratized to anyone with great ideas and creativity at sap we are at the Forefront of business AI sap business AI is about achieving real business results for the Enterprise and making AI relevant reliable and responsible in tomorrow's day two keynote we'll show you how we expend embed AI capabilities into our Solutions and how Jewel the generative AI co-pilot embedded across our applications truly understands your business we are relentlessly working to bring these use cases to our customers to facilitate this we have been educating sap employees on best practices and providing AI tools internally to accelerate gen embedded product development for today I want to talk about how we make these learnings available to you btp AI Foundation is your way to build relevant reliable and responsible AI infused applications and extensions yourselves it includes everything you need to run your business ready AI in these applications that you build on S on saps btp from ready to use AI services to Ai runtimes and life cycle management we provide the tooling for the management of generative AI capabilities and ensure business data connectivity this is the same Foundation we use ourselves to embed AI into our products and we designed it with security governance and Trust in mind from the ground up and with Enterprise grade as you would expect and demand from all sap offerings so I'm sure you're now asking yourself how do I access these large language models as part of the AI Foundation I am pleased to announce the release of our generative AI hub yes plan for General availability by December the Hub provides developers instant access to Azure open Ai and Falcon B Falcon 40b language models and early next year we're going to add more models including metas Lama 2 and LF Alpha with this access developers can experiment by submitting a prompt to multiple large language models to compare the generated outcomes and identify the best student model for the task furthermore they can then power Mission critical processes with complete control and transparency with features like built-in prompt history all delivered with Enterprise grade security and data privacy generative AI Hub also provides tooling for developers to supercharge your gen development processes you'll be able to leverage grounding capabilities based on Enterprise data to ensure the proper context such as leveraging Hana Cloud Vector store as shown by Sabrina earlier today this for instance enables retrieval augmented generation or RG rag uh a technique to extract the most relevant information for context aware prompting of large language models let's look at how this works imagine you a developer named Sunil who is eager to explore a new generative AI Hub with rag grounding capabilities normally this would be timec consuming but with our userfriendly UI s can save a lot of time to continue from his last session soil starts in the prompt management section where he was experimenting with grounding he starts this session by filtering on his save prompts selecting the grounding option in the collections to find something he did earlier this takes him straight to what he was looking for and with a simple click of opening editor button he changes the prompt for what he wants today which is how do you calculate the sorage path in a hammer Hanah golf engine Falcon 40b is already selected for this model for this run so the next thing he needs to do is choose a grounding option which in this case is the Hanah Cloud graph documentation now he executes the workflow what's happening in the background the query goes into the hand of VOR Vector store whose output in turn is the input grounding into Falcon looking at the results soil finds a good response but you know he was actually hoping for a work working code snippet this is where the strength of the generative AI Hub shines soel can select from multiple providers and versions of large language models here he decides to use gp4 since he's seen that gives more
2023-11-09 19:04
