On the SAP TechEd 2023 stage, Walter Sun, Global Head of AI at SAP, unveiled AI Foundation, a one-stop-shop for developers to bootstrap their own Business AI solutions by creating AI- and generative AI-powered extensions and applications on SAP Business Technology Platform (SAP BTP).
AI Foundation is SAP’s all-in-one AI toolkit, offering developers AI that’s ready-to-use, customizable, grounded in business data, and supported by leading generative AI foundation models. It is also the basis for AI capabilities that SAP embeds across its portfolio.
Let’s delve into each bucket of this complete set of services to better understand how you can build relevant, reliable and responsible AI-infused applications and extensions yourself starting today, with security, governance and trust in mind.
Let’s start with the first layer of the stack, AI services. These are AI models pre-trained on business-relevant data to extend your own applications, covering a wide range of business use cases, including:
1/ Document Processing to reduce manual effort and costs with Document Information Extraction. This service helps you to process large amounts of business documents that have content in headers and tables. You can use the extracted information, for example, to automatically process payables, invoices, or payment notes while making sure that invoices and payables match. After you upload a document file to the service, it returns the extraction results from header fields and line items. Using this service, SAP customers have processed more than 350 million documents since the beginning of 2023.
Starting from December 2023, we will supercharge the existing service with generative AI capabilities and launch Document Information Extraction, premium edition, supporting over 40+ languages with easy extensions for any kind of document.
I strongly recommend watching this 20-min session from SAP TechEd 2023 including real-world examples from our product experts or the shorter demo below to see it in action:
2/ Recommendations to enhance human-decision making, with Personalized Recommendation and Data Attribute Recommendation services.
Personalized Recommendation gives visitors to your website highly personalized recommendations (next-item, similar-item, smart-search and user-affinity) based on their browsing history and/or item description to elevate user experiences. For instance, SAP is leveraging these capabilities to process more than 4.2 million learning recommendations for SAP SuccessFactors customers each month. This deep dive session including live demos with our product expert will give you everything you need to get started.
Data Attribute Recommendation uses free text, numbers and categories as input to classify entities such as products, stores and users into multiple classes and to predict the value of missing numerical attributes in your data records. This service helps you to speed up data management processes and increase data consistency and accuracy.
New capabilities introduced at SAP TechEd allow data scientists to train models faster, process multiple inference requests offline, and more easily identify parameters that influence predictions, improving process efficacy and prediction accuracy. New explainability capabilities also help users better understand how models use data.
Watch this demo to understand how SAP embeds Data Attribute Recommendation into SAP S/4HANA Cloud to auto-complete sales orders.
3/ Machine Translation to translate content into multiple languages with SAP Translation Hub.
This service helps you speed up the translation of software texts (e.g. user interfaces) and related documents – with high quality and accuracy. You can use a repository of SAP-approved translations and terminology as well as machine translation, which has been trained with focus on SAP-specific content. If desired, you can reuse your existing UI translations as well. You can access the translation resources by consuming a range of API methods or by using integrated workflow scenarios on a UI. It is also part of other SAP software offerings, such as SAP Enable Now.
If you prefer to have a new AI model to work on from scratch, our AI workload management offering provides all the necessary to create models, train them, evaluate their accuracy and publish them for inferencing.
SAP AI Core offers capabilities to confidently deploy and integrate your AI models designed for SAP applications, cost-efficiently at scale while preserving privacy and compliance. It enables you to execute pipelines, serve inference requests, benefit from multitenancy support, productize your AI content and expose it as a service to consumers in the SAP BTP marketplace.
SAP AI Launchpad enables you to transparently manage your AI models. It allows you to connect to multiple AI runtimes, including SAP AI Core, and centralize AI lifecycle management for your AI scenarios with a convenient user interface. You can monitor model performance statistics continuously and retrain as needed.
In this example, see how a company managed to reduce its CO2 footprint using a generative AI-powered application on SAP BTP that interacts with Microsoft Azure AI through SAP AI Core and SAP AI Launchpad.
To fast-track your generative AI development, AI Foundation offers the generative AI hub, giving instant access to a broad range of large language models (LLMs) from different providers, such as GPT-4 by Azure OpenAI or OpenSource Falcon-40b. With this access, you’ll be able to orchestrate multiple models, whether programmatically via SAP AI Core or via the playground within SAP AI Launchpad.
The generative AI hub provides tooling for prompt engineering, experimentation, and other capabilities to accelerate the development of your SAP BTP applications infused with generative AI, in a secure and trusted way. AI development teams can submit a prompt to multiple LLMs, compare the generated outcomes to identify the best-suited model for the task, and gain greater control and transparency with the built-in prompt history.
Read this blog post from our product manager to learn more or start the tutorial right away.
With SAP HANA Cloud, you can ground AI with your unique business data and context using the vector engine and similarity search. SAP HANA Cloud’s AI function libraries (Predictive Analysis Library, Automated Predictive Library) allow to implement classification, regression or time series forecasting scenarios applied directly to your business data, which can be orchestrated by SAP AI Launchpad.
In addition, the new generative AI hub in SAP AI Core introduced above will connect to the new vector capabilities in SAP HANA Cloud (available early 2024), helping developers reduce model hallucinations and incorporate contextual data as embeddings (or groundings) to deliver more tailored results to specific use cases.
A vector datastore manages unstructured data - such as text, images, or audio - in high-dimensional vectors or embeddings, to provide long-term memory and better context to AI models. This makes it easy to find and retrieve similar objects quickly, for example, by asking a question using natural language. This both simplifies interactions with LLMs and empowers developers to securely implement generative AI in applications.
SAP HANA Cloud will natively store and search vector embeddings, which are numerical representations of objects, along with business data as part of its industry-leading multi-model processing capabilities to power intelligent data applications. With these vector capabilities, SAP HANA Cloud will enable RAG, facilitating the combination of LLMs with private business data.
For Data Management, SAP Datasphere, a multicloud data fabric solution that enables both hybrid and cloud native architectures, leverages AI by allowing you to integrate your data and AI platforms to deliver data in business terms from multiple systems and locations.
SAP Datasphere includes AI in several ways: data integration, data quality & cleansing (such as correcting missing values or duplicates), data governance, analytics (such as detecting trends, analyze historical data, and make predictions), and natural language processing.
SAP partners with the leading general-purpose AI vendors and LLMs providers to help ensure that SAP customers keep up with the fast pace of innovation and have the flexibility they need.
To get the most value in traditional business areas such as finance, sales, or supply chain, our customers require foundation models designed for business context. SAP will fine-tune generic large language models on SAP anonymized data as well as create proprietary foundation models based on our vast structured business data. These models will be able to address questions we face every day in business that large language models cannot, such as predicting invoice payment dates and supplier delivery quality or proposing efficiency improvements to a business process.
Stay tuned for more information in 2024. In the meantime, I’d strongly recommend watching this insightful SAP TechEd session on the topic:
To sum up, AI Foundation includes everything developers need to create business-ready AI applications, from ready-to-use AI services and access to the top large language models to vector database capabilities and AI runtime and lifecycle management.
▶︎ Get started with AI Foundation on SAP.com now, by selecting the capability that fits most to your business needs.
▶︎ Interested in our upcoming product innovations? Check out the AI Foundation roadmap.
▶︎ Engage with our community through blog posts and Q&A on the overall AI SAP Community page. I especially recommend this blog post series that dives deeper into various aspects of SAP’s approach to generative AI and its technical underpinnings.
AI Foundation is SAP’s all-in-one AI toolkit, offering developers AI that’s ready-to-use, customizable, grounded in business data, and supported by leading generative AI foundation models. It is also the basis for AI capabilities that SAP embeds across its portfolio.
Let’s delve into each bucket of this complete set of services to better understand how you can build relevant, reliable and responsible AI-infused applications and extensions yourself starting today, with security, governance and trust in mind.Ready-to-use AI services
Let’s start with the first layer of the stack, AI services. These are AI models pre-trained on business-relevant data to extend your own applications, covering a wide range of business use cases, including:
1/ Document Processing to reduce manual effort and costs with Document Information Extraction. This service helps you to process large amounts of business documents that have content in headers and tables. You can use the extracted information, for example, to automatically process payables, invoices, or payment notes while making sure that invoices and payables match. After you upload a document file to the service, it returns the extraction results from header fields and line items. Using this service, SAP customers have processed more than 350 million documents since the beginning of 2023.
Starting from December 2023, we will supercharge the existing service with generative AI capabilities and launch Document Information Extraction, premium edition, supporting over 40+ languages with easy extensions for any kind of document.
I strongly recommend watching this 20-min session from SAP TechEd 2023 including real-world examples from our product experts or the shorter demo below to see it in action:
2/ Recommendations to enhance human-decision making, with Personalized Recommendation and Data Attribute Recommendation services.
Personalized Recommendation gives visitors to your website highly personalized recommendations (next-item, similar-item, smart-search and user-affinity) based on their browsing history and/or item description to elevate user experiences. For instance, SAP is leveraging these capabilities to process more than 4.2 million learning recommendations for SAP SuccessFactors customers each month. This deep dive session including live demos with our product expert will give you everything you need to get started.
Data Attribute Recommendation uses free text, numbers and categories as input to classify entities such as products, stores and users into multiple classes and to predict the value of missing numerical attributes in your data records. This service helps you to speed up data management processes and increase data consistency and accuracy.
New capabilities introduced at SAP TechEd allow data scientists to train models faster, process multiple inference requests offline, and more easily identify parameters that influence predictions, improving process efficacy and prediction accuracy. New explainability capabilities also help users better understand how models use data.
Watch this demo to understand how SAP embeds Data Attribute Recommendation into SAP S/4HANA Cloud to auto-complete sales orders.
3/ Machine Translation to translate content into multiple languages with SAP Translation Hub.
This service helps you speed up the translation of software texts (e.g. user interfaces) and related documents – with high quality and accuracy. You can use a repository of SAP-approved translations and terminology as well as machine translation, which has been trained with focus on SAP-specific content. If desired, you can reuse your existing UI translations as well. You can access the translation resources by consuming a range of API methods or by using integrated workflow scenarios on a UI. It is also part of other SAP software offerings, such as SAP Enable Now.
AI workload management
If you prefer to have a new AI model to work on from scratch, our AI workload management offering provides all the necessary to create models, train them, evaluate their accuracy and publish them for inferencing.
SAP AI Core offers capabilities to confidently deploy and integrate your AI models designed for SAP applications, cost-efficiently at scale while preserving privacy and compliance. It enables you to execute pipelines, serve inference requests, benefit from multitenancy support, productize your AI content and expose it as a service to consumers in the SAP BTP marketplace.
SAP AI Launchpad enables you to transparently manage your AI models. It allows you to connect to multiple AI runtimes, including SAP AI Core, and centralize AI lifecycle management for your AI scenarios with a convenient user interface. You can monitor model performance statistics continuously and retrain as needed.
In this example, see how a company managed to reduce its CO2 footprint using a generative AI-powered application on SAP BTP that interacts with Microsoft Azure AI through SAP AI Core and SAP AI Launchpad.
Generative AI management
To fast-track your generative AI development, AI Foundation offers the generative AI hub, giving instant access to a broad range of large language models (LLMs) from different providers, such as GPT-4 by Azure OpenAI or OpenSource Falcon-40b. With this access, you’ll be able to orchestrate multiple models, whether programmatically via SAP AI Core or via the playground within SAP AI Launchpad.
The generative AI hub provides tooling for prompt engineering, experimentation, and other capabilities to accelerate the development of your SAP BTP applications infused with generative AI, in a secure and trusted way. AI development teams can submit a prompt to multiple LLMs, compare the generated outcomes to identify the best-suited model for the task, and gain greater control and transparency with the built-in prompt history.
Read this blog post from our product manager to learn more or start the tutorial right away.
Generative AI Hub Playground Demo
Business Data & Context
With SAP HANA Cloud, you can ground AI with your unique business data and context using the vector engine and similarity search. SAP HANA Cloud’s AI function libraries (Predictive Analysis Library, Automated Predictive Library) allow to implement classification, regression or time series forecasting scenarios applied directly to your business data, which can be orchestrated by SAP AI Launchpad.
In addition, the new generative AI hub in SAP AI Core introduced above will connect to the new vector capabilities in SAP HANA Cloud (available early 2024), helping developers reduce model hallucinations and incorporate contextual data as embeddings (or groundings) to deliver more tailored results to specific use cases.
A vector datastore manages unstructured data - such as text, images, or audio - in high-dimensional vectors or embeddings, to provide long-term memory and better context to AI models. This makes it easy to find and retrieve similar objects quickly, for example, by asking a question using natural language. This both simplifies interactions with LLMs and empowers developers to securely implement generative AI in applications.
SAP HANA Cloud will natively store and search vector embeddings, which are numerical representations of objects, along with business data as part of its industry-leading multi-model processing capabilities to power intelligent data applications. With these vector capabilities, SAP HANA Cloud will enable RAG, facilitating the combination of LLMs with private business data.
For Data Management, SAP Datasphere, a multicloud data fabric solution that enables both hybrid and cloud native architectures, leverages AI by allowing you to integrate your data and AI platforms to deliver data in business terms from multiple systems and locations.
SAP Datasphere includes AI in several ways: data integration, data quality & cleansing (such as correcting missing values or duplicates), data governance, analytics (such as detecting trends, analyze historical data, and make predictions), and natural language processing.
Foundation Models
SAP partners with the leading general-purpose AI vendors and LLMs providers to help ensure that SAP customers keep up with the fast pace of innovation and have the flexibility they need.
To get the most value in traditional business areas such as finance, sales, or supply chain, our customers require foundation models designed for business context. SAP will fine-tune generic large language models on SAP anonymized data as well as create proprietary foundation models based on our vast structured business data. These models will be able to address questions we face every day in business that large language models cannot, such as predicting invoice payment dates and supplier delivery quality or proposing efficiency improvements to a business process.
Stay tuned for more information in 2024. In the meantime, I’d strongly recommend watching this insightful SAP TechEd session on the topic:
Supercharge Your Development with AI Foundation Now
To sum up, AI Foundation includes everything developers need to create business-ready AI applications, from ready-to-use AI services and access to the top large language models to vector database capabilities and AI runtime and lifecycle management.
▶︎ Get started with AI Foundation on SAP.com now, by selecting the capability that fits most to your business needs.
▶︎ Interested in our upcoming product innovations? Check out the AI Foundation roadmap.
▶︎ Engage with our community through blog posts and Q&A on the overall AI SAP Community page. I especially recommend this blog post series that dives deeper into various aspects of SAP’s approach to generative AI and its technical underpinnings.
- SAP Managed Tags:
6 Comments
