Skip to Content

Reimagining customer service with generative AI and SAP HANA Cloud: A story of empowerment

Steven Dierick
Jul 11, 2024

In today’s digital age, customers expect personalized, efficient, and cost-effective support from their service providers. As an architect and SAP Consultant at Capgemini, I am thrilled to share a proof of concept that has the potential to transform customer service experiences. Our team has developed a cutting-edge chatbot that leverages the power of generative AI, SAP HANA Cloud, and the Langchain framework to provide personalized and efficient support to customers.

Let’s explore how this technology empowers customers through the story of Emma, a homeowner with questions about her water supplier. One evening, Emma is preparing to move out of her current residence and seeks clarification on what to do when changing water companies. Rather than calling customer service or browsing through FAQ sections, she turns to the company’s chatbot integrated with SAP HANA Cloud. Emma types her question into the chat interface: “What do I need to do when changing water companies when I move out?”

Behind the scenes, the process unfolds with seamless efficiency. Emma’s question is first converted into a digital format that the system can understand, known as “question embedding.” This process transforms her text into a numeric vector, encapsulating the essence of her inquiry in a language the system can process.

Once Emma’s question is embedded, the SAP HANA Cloud Vector Engine springs into action. It performs a similarity search across a vast database of text documents, including guides, past customer queries, and water usage tips—all of which have been similarly transformed into numeric vectors.

The Vector Engine quickly identifies the most relevant documents and text chunks that match Emma’s query, pinpointing the information that could explain the process of changing water companies when moving out. This could be related to notifying the current water provider, setting up a new account with the new provider, or understanding any potential fees associated with the change.

These best matches are then analyzed by a large language model, a sophisticated AI that understands and processes natural language. The model evaluates the content, context, and nuances of the matched text chunks, synthesizing an accurate and coherent response.

Back on her couch, Emma receives a prompt reply from the chatbot on her screen. The chatbot explains the step-by-step process of changing water companies when moving out, including the necessary notifications, paperwork, and potential fees. It also offers tips for ensuring a smooth transition and avoiding any disruptions to her water service.

This chatbot, empowered by the SAP HANA Cloud Vector Engine, not only provides Emma with quick and precise answers but also enhances her confidence and satisfaction with her utilities provider. The technology ensures that each customer interaction is handled with care, providing personalized and informed responses directly through an accessible, user-friendly chat interface, anytime and anywhere.

But what makes this solution truly revolutionary is the SAP HANA Cloud Vector Engine. This powerful AI-enabled database enables fast and efficient processing of large datasets, allowing us to store and analyze vast amounts of customer data. The Vector Engine is built on top of SAP HANA’s in-memory computing capabilities, which enables it to process complex algorithms and machine learning models at incredible speeds.

One of the key features of the Vector Engine is its ability to support Retrieval-Augmented Generation (RAG) architecture. RAG is a method that combines the benefits of retrieval-based models and generative models for machine learning. In the context of our chatbot, RAG allows us to split business documents into chunks, each represented by an embedding. An embedding is a vector that is a numerical representation of a text chunk, capturing its meaning and context.

When a user asks a question, our chatbot uses the Vector Engine to perform a similarity search, comparing the user’s question with the embeddings in the vector database. This allows the chatbot to retrieve the most relevant documents and generate a response that is accurate and contextually appropriate.

For our clients, this solution offers a range of benefits, including improved customer satisfaction, reduced operational costs, and enhanced customer insights. By automating routine tasks and providing instant responses, our chatbot reduces the need for human intervention, resulting in significant cost savings. Additionally, our chatbot provides valuable insights into customer behavior, allowing our clients to inform business decisions and drive growth.

In conclusion, our proof of concept demonstrates the potential of generative AI and SAP HANA Cloud to revolutionize customer service experiences. By leveraging the Vector Engine and RAG architecture, we can create personalized, efficient, and cost-effective support models that drive customer satisfaction and loyalty. We believe that this solution has the potential to transform the customer service landscape, and we’re excited to explore its possibilities with our clients.

Would you like to learn more? Contact our expert!

Author

Steven Dierick

Senior Manager, Business Intelligence and AI
Meet Steven Dierick, a seasoned professional with nearly two decades of experience at Capgemini. Starting his career as a developer, Steven quickly demonstrated his expertise and passion for technology, transitioning to a data engineer role specializing in SAP Analytics after just three years. Today, he is at the forefront of innovation in artificial intelligence, applying his deep technical background and extensive experience to drive transformative AI solutions. Steven’s unique blend of development skills and analytical prowess makes him a key player in leveraging AI technologies to deliver impactful results.