bg

Transform Your Supply Chain Planning and Marketing Strategies with Google Cloud and SAP Integration

Cloud Marketplace

Explore Pi Agent pi-logo

Enable Generative AI Capabilities with SAP Datasphere and Google Cloud

July 11, 2023 | Asheesh Gupta

Blog / Enable Generative AI Capabilities with SAP Datasphere and Google Cloud

When ChatGPT rose to stardom, it wasn’t just a technological feat—it was a management lesson in data literacy. This unexpected star demonstrated two significant things. First, with enough high-quality data, learning models could produce equally high-quality outputs. Second, it highlighted that poor-quality data could lead to substantial gibberish.

One of the key learnings from the ChatGPT episode was the glaring need for simplifying data analysis in the enterprise market. Enterprises needed a way to transform raw data into valuable insights without the hassles of complex SQL queries or setting up machine learning models. 

Although ChatGPT was made available to the public earlier, there’s no denying that Google is a pioneer and leader in Generative AI technology. Through Google Cloud’s Vertex AI, it is now opening up state-of-the-art Large Language models like PaLM 2, trained on billions of parameters, and the entire Generative AI ecosystem to enterprises securely.

Similarly, it is an undeniable fact that when it comes to quality enterprise data, it is hard to beat SAP. With millions of organizations running their critical business processes on SAP for decades, the SAP data landscape has evolved remarkably. Now, enterprises can seamlessly integrate their SAP ECC/ SAP S/4HANA and other SAP LoB data through SAP BTP/Datasphere’s native integration capabilities.

Through the recent partnership announced at Sapphire, Google and SAP have come together to offer the joint capability of SAP’s solid data foundation through Datasphere and state-of-the-art Generative AI Capabilities through Vertex AI. This has generated significant excitement in the enterprise world, as enterprises recognize that both a solid data foundation and Generative AI capabilities are critical to leveraging the benefits offered by generative AI technology.

While SAP brings in the needed enterprise data at one place in Datasphere through built-in native connectors, GenAI in Vertex AI simplifies the process for developers and data scientists to access, customize, and deploy foundation models. It offers a variety of tools, automated workflows, and starting points. 

Now using the below architecture of joint SAP BTP and GCP services, organizations can realize various use cases across supply chain, marketing, sales, and finance. They can in real time identify bottlenecks in their Supply Chain network by blending data from multiple data sources both structured and unstructured. 

What kind of data? 

  • The data can be first party data streaming live from field or stored in various databases , ERPs and document repositories of enterprises
  • Second-party data available through partner networks or,
  • Third-party data available for free through the internet, 250+ datasets stored in Google Bigquery or paid datasets from agencies or various analysts. 

The end goal is to provide insights and capabilities to enterprises that were not possible earlier or were available with a lot of effort, costs, and delays associated. With this combined GCP and BTP architecture, enterprises have all Generative AI capabilities with the prerequisite of a solid data foundation. 

What kind of impact does it have? 

  • Now, enterprises can have better control over their supply chain by receiving proactive alerts on which purchase orders are impacted with delays introduced by port congestion, cyclones and hurricanes, transport strikes or other unplanned events.
  • Financial impacts are immediately available due to these delays and recommendations are available in real-time to minimize the impact of these unforeseen events in the form of auto-generated emails sent to other approved vendors requesting quotations.
  • In product development, auto-generated product designs are available for designers to choose from or further enhance based on market demand and margins from past sales. Customers can search the  right products as per their taste assisted by generative AI bots.

This architecture diagram depicts the integration between SAP BTP and Google Cloud services for effective data handling and Generative AI deployment 

The architecture comprises the following services from SAP BTP and Google Cloud: 

  1. SAP Datasphere : Brings together multiple data sources like SAP S4 HANA and other SAP Line of Business products, Google Bigquery Data sources as well as other databases and second and third-party structured data sources.
  2. GCP Vertex AI Platform : Hosts all of Google’s state of art foundation models for Large Language models (LLMs) like PalM text, chat, imagegen etc. and a model garden that hosts other open source models. GenAI services in Vertex AI simplifies the process for developers and data scientists to access foundational large language data models, perform prompt engineering, fine-tune using techniques like parameter-efficient fine-tuning, and create and deploy ML pipelines.
  3. Fed ML Python Library in Vertex AI : This enables federated AI/ML capabilities which makes sure that Large Language models can access data stored in Datasphere without physically moving data from Datasphere to Vertex AI.
  4. Node.js-based applications hosted on SAP BTP : Node.js-based pre-built applications already available in SAP BTP Store for managing workflows and visualization.
  5. Other platform capabilities : Leverages capabilities like security, data governance, MDM, and data compliances, already pre-built in SAP BTP and GCP platforms. 

How Pluto7’s Data Platform ‘Planning in a Box’ Leverages Google Cloud’s GenAI Foundation Models

At Pluto7, we have leveraged the power of Google Cloud’s GenAI to enhance our data solutions. We have taken Google Cloud’s foundation models and customized them to cater to various use cases across supply chain, marketing, sales, and finance.

Our process begins by understanding the specific needs and challenges of the enterprise. Then we fine-tune Google Cloud’s foundation models using enterprise data to create solutions that are uniquely suited to address those challenges.

For instance, to make demand planning more responsive and accurate, we use GenAI to integrate over 250 external demand signals into our models. This not only broadens our forecasting perspective but also enhances the precision of our predictions.

Similarly, to facilitate interaction with data, we use GenAI to create a chat-like interface. This interface allows users to pose specific questions, which our models then convert into SQL queries.

The customizability of Google Cloud’s GenAI foundation models and our expertise in fine-tuning these models enable us to offer solutions that are highly relevant, efficient, and valuable to our customers. It’s a combination that not only enables enterprises to realize the benefits of GenAI but also to do so in a way that is specifically tailored to their needs.

What makes Pluto7’s Data Platform unique?

Pluto7’s Data Platform is the backbone of our sophisticated data-driven solutions, and its uniqueness lies in its capability to integrate diverse data sources seamlessly, both from within and outside the SAP environment. It can effortlessly incorporate data from Salesforce, external data sources like Google Trends, and even partner data from vendors, creating a broad data canvas for businesses to draw insights from.

It is made possible by the power of the Google Cloud Cortex Framework, which works behind the scenes to create a coherent and structured view of all your data on Google Cloud. This integrated perspective of data, referred to as the ‘data foundation’, is the key that unlocks the potential of AI and Generative AI. Without a solid data foundation, the full capabilities of AI cannot be realized.

Upon this foundation, the platform leverages Vertex AI’s Generative AI capabilities to extract valuable insights and generate responses. These insights, gleaned from the structured view of your data, are presented via a chat-like interface on Pluto7’s data platform. This ensures that interacting with data becomes an intuitive process, regardless of a user’s technical expertise.

In essence, Pluto7’s Data Platform is the meeting point of different data streams and state-of-the-art AI technologies, all working together to facilitate easy interaction with data, generating valuable insights, and aiding swift, intelligent decision-making. By focusing on integrating disparate data sources and creating a stable data foundation, our platform sets the stage for the magic of Generative AI to unfold.

Unleashing the Potential of Generative AI for SAP Users

Thanks to SAP’s robust partnership with Google Cloud, SAP users can tap into Google Cloud’s cutting-edge solutions with ease and confidence. This symbiotic relationship ensures that SAP users can securely access and deploy the groundbreaking capabilities of Generative AI, while minimizing integration hassles and maximizing value.

At the heart of this integration is Pluto7, which brings Google Cloud’s Generative AI capabilities right into your SAP landscape. Our solutions operate within Datasphere, leveraging Google Cloud’s Generative AI tools to enhance SAP data in a secure, seamless, and impactful manner through SAP Business Technology Platform (BTP).

As a result, SAP users can engage with Generative AI without needing to step out of their familiar SAP environment. Whether it’s improving demand forecasting with external signals, dynamically writing SQL queries to adapt to changing business scenarios, generating content on-demand, or effortlessly training custom LLM models, Pluto7’s solutions offer plug-and-play capabilities that translate into immediate benefits for SAP users.

I understand that all this information might feel overwhelming. So, let’s break it down. We are offering an exclusive Pluto7 Workshop Powered by Google Cloud, where we will work with you to launch a pilot, delivering actionable results in 4 weeks or less. If you’re having conversations internally about integrating Generative AI with your SAP data, this is the perfect platform to explore, learn, and drive transformation.

Very few companies globally have truly harnessed the power of their data. At Pluto7, we’ve mastered this, and we want to share this competitive advantage with you. Are you ready to redefine your company’s data strategy and set new industry benchmarks?

ABOUT THE AUTHOR

Asheesh Gupta is the Head of Enterprise Architecture at Pluto7, bringing over 20 years of experience in technology consulting, system integration, and service delivery. He excels in architecting and delivering data-driven solutions on hyperscalar cloud platforms like GCP and SAP BTP. With a focus on supply chain optimization through digital transformation and AI, Asheesh’s work at Pluto7 continues to drive value and innovation in enterprise technology.

Connect with Asheesh on LinkedIn