SAP TechEd 2023: What’s new in AI and Data & Analytics

Both Reporting and AI need correct and consistent data, so let’s start by looking at SAP's Data Warehousing solutions. Next to the traditional SAP BW environment and the underlying HANA Cloud Database, SAP recently made a switch to the cloud-based Data Warehouse solution SAP Datasphere, on which there was a lot of emphasis during TechEd. Although SAP BW will still be licensed for quite a while, BW 7.5 will be supported until 2027 and BW/4HANA until 2040, we see that the majority of new functionalities are exclusively focused on SAP Datasphere. Datasphere is positioned as the solution to realize a real-time Business Data Fabric in which you can manage all business-critical data and overcome the main challenges that exist when driving value from Data. Openness and self-service are at the core of the Datasphere concept, with SAP having established partnerships with (sometimes) competing vendors such as Databricks, Google and Collibra. The increasingly closer integration between SAP Analytics Cloud and Datasphere illustrates the move to a business-friendly data environment SAP is striving for. SAP will very much focus on further bridging the gap between technology and business by implementing a couple of new features.

The most important one mentioned, which will become available in Q4 2023 is the Semantic Onboarding (see image below). This feature will leverage all the metadata that is available in your SAP source systems and make this available in SAP Datasphere. By doing so, self-service BI by business users will become a lot easier since the data is directly shown in fields that are recognizable to business users instead of showing technical field names. This new functionality will be generally available in Q4 2023.

Next to this there is a huge emphasis on the low code Analytic Model that was introduced in May this year. This is the place where all modeling comes together in a way that data can be optimally consumed by external tools. Self-service business users can create their own data-models here in such a way that data consistency and proper authorization is taken care of by re-using IT managed artifacts. The Analytic Model will also be the most important object in Datasphere for communication to SAC, with more advanced functionalities such as (exception) aggregation, multi-facts and support for external hierarchies and stacking, all planned for the (near) future.

Other announcements for Datasphere include:

• Creation of folders in the Data Builder to better structure your artifacts.
• Version control in the form of Compatibility Contracts.

Enhanced partnerships to create a more open ecosystem with:

* Databricks - Further investment in integrating dataflows.
* Confluent - Event streaming.
* Collibra - Combining Non-SAP and SAP Data catalog information.
* Data Robot - Self-service AI/ML scenarios.
* Google Cloud - Data center (next to AWS and Microsoft).

• The Business Builder will remain available as a slim layer between Data Builder and Analytic Models. This will become effective in 2024.

• Data Intelligence functionality will be gradually integrated towards Datasphere, with a primary focus on Machine Learning and complex pipelining functionalities to be made available in 2024.
  

With all the fuss around Datasphere it can sometimes be easy to forget SAP BW, for which SAP is still heavily investing in the BW Bridge. The Bridge allows integration and staging of BW data in Datasphere through a BW instance that runs on the same HANA Cloud-engine as Datasphere itself. This allows customers to move to the public cloud while still leveraging their existing investment and skills in SAP BW and ABAP technologies. Moreover, it also allows customers to maintain as much of their BW-based business semantics as possible. The various scenarios on how to integrate Datasphere with SAP BW over time are listed below. Read more about SAP’s Business Data Fabric and BW Bridge (and possible migration scenarios) in our whitepaper.

The game-changing update to HANA Cloud’s multi-model capabilities this year are the so-called vector capabilities. Through this new multi-dimensional Vector Engine, HANA can derive context from unstructured data. This basically means that HANA will be able to manage images, social media posts or audio, as vector embeddings. The idea here is to allow HANA developers to leverage large language models within related data applications. This update is currently scheduled to be released in Q1 of the next year.

Other innovations in the SAP HANA area include:

• Better price/performance option for customers, allowing them to configure a more cost-optimized environment.
• Self-service migration to support on-premise to cloud migration.
• Intelligent data-tiering by determining which objects can be off-loaded to disk.

Having a proper foundation in SAP Datasphere and Hana Cloud allows for getting better insights in the Reporting & Analytics area with SAP Analytics Cloud (SAC). When it comes to story development there is going to be a lot of focus on the Unified Story mode (aka Optimized mode). SAP is actively asking customers to migrate from Classic Story mode to the Unified Story mode to benefit from new functionalities such as the Lightweight Viewer, which is set to improve Story performance by up to 63%. This functionality will be released in a controlled way starting from Q1 2024 and will not be made available to Classic Stories. This change will also lay the foundation for composable BI Enhancements that will come in the first half of 2024, by integrating the Digital Boardroom feature in the Unified Story and allowing the integration of custom widgets to the Story. A fun note regarding those custom widgets: SAP has released a nice hands-on guide on how to implement them, which you can find by clicking here. Looking at this roadmap, it is our advice to start migrating Classic Stories to Unified Stories as soon as possible in order to prepare for these new capabilities.

SAP is also investing in AI tooling for reporting by further developing the 'Just Ask' feature, which applies generative AI to search-driven analytics. SAP’s large language models enhance answer accuracy and consistency, surpassing existing search-driven solutions in Q1 2024. SAC’s planning capabilities will be expanded through a unified model that combines both its planning and analytical facets, also around Q1 of next year. Further down the road, functionality will become available that allows advanced simulations that can be integrated to adjust planning. This is currently in a pilot phase with a limited number of customers but certainly worth keeping a close eye on.

Continuing the trend of implementing AI wherever possible, we see SAP expanding its AI Core application with the generative AI hub (see image below). This hub gives developers direct access to a multitude of large language models such as Falcon-40b and Azure OpenAI, with more models such as Llama2 and Aleph Alpha (you might recognize this name as one of the companies in which SAP invested in the summer of this year) being added in 2024. Apart from the BTP-widespread implementation of AI co-pilot Joule, SAP’s further AI-related plans are to provide templates for prompt management, support for libraries and even proprietary tabular transformer-based models. Business AI based on foundation models will also continue to grow and become increasingly present across SAP’s business solutions. Overall, it seems to us that what SAP rather lacked in terms of commitment to AI and machine learning in the past years, it is doubling down on in 2023... even if it sometimes comes at the expense of existing (in some cases long-awaited and now postponed) roadmap items.

SAP BO 4.2 will reach its end-of-priority support by the end of next year, so many customers will have to move to version 4.3 before then. If we look ahead, we see SAP’s new BusinessObjects 2025 release coming up in the fourth quarter of 2024, with annual support packages following that release. In practice, this means that BO users that stay current with the latest version will enjoy mainstream support from Walldorf up to 2030. Note that by saying ‘BO users’, SAP includes Web Intelligence (WeBi), single source .unx universes, Analysis for Microsoft Office (AfO), Lumira Designer, Crystal Reports and the BusinessObjects BI platform. Solaris and AIX will no longer be supported after SAP BO 4.3.


Finally, along with these BO updates, SAP is also planning to launch SAP Data Services 2025 and SAP Information Steward 2025 as successors to their respective 4.3-versions, in Q4 of 2024. Both solutions will be available either in the Cloud or on-premise. You can find more details on this announcement here.

Credits

This blog about the highlights of SAP TechEd 2023 was written by our experts Rogier Schipper and Lars van der Goes.