Transforming Customer Care with Generative AI

AI Generativa - Intellico


In today’s competitive business environment, leveraging innovative technologies to enhance customer care is not just an option—it’s a necessity. Generative AI stands at the forefront of this transformation, offering significant benefits, diverse applications, and strategic advantages, albeit with certain risks.

According to research by Osservatori Artificial Intelligence (Politecnico di Milano) for most of the companies with active AI projects, Generative AI has led to an acceleration of the adoption path. Moreover, about 3 out of 5 large companies are starting to reflect on and work on applications and projects involving Generative AI(1).

Moreover an analysis by McKinsey(2) highlighted that customer operations are among the four uses cases where Generative AI could deliver about 75 percent of the value.

Why Generative AI in Customer Care?

Generative AI technologies have revolutionized how businesses interact with customers, delivering personalized, efficient, and scalable solutions. The main applications range from automated customer support and personalized recommendations to feedback analysis and proactive service alerts. The key benefits include:

  • Enhanced Customer Experience: AI-driven chatbots provide timely, accurate, and personalized responses, elevating the customer service experience;
  • Operational Efficiency: automating routine inquiries allows staff to focus on complex issues, improving service quality while reducing operational costs;
  • 24/7 Availability: Generative AI ensures customers receive immediate support at any time, significantly improving satisfaction and engagement.

Several academics and companies are trying to analyze the long-term impact on productivity highlighting significant outcome. For instance, a research found that at one company with 5,000 customer service agents, the application of generative AI increased issue resolution by 14% an hour and reduced the time spent handling an issue by 9%(3).

However, businesses must navigate certain risks, such as the potential for generating inaccurate or irrelevant responses. How to solve this limitation?

In Intellico, we have introduced in our generative solutions two technologies: Retrieval Augmented Generation (RAG) and Knowledge Graph (KG).

The combination of these two technologies has significantly improved the performance of the generative chatbot, achieving excellent levels of B2C response accuracy even in contexts where the adherence to sequences and procedures is critical, such as suggesting recipes from specialists and providing technical instructions to the end customer.

Retrieval Augmented Generation (RAG): how it works

Retrieval Augmented Generation (RAG) integrates the power of knowledge retrieval with advanced natural language generation. It operates in 2 steps:

  • Information Retrieval: dynamically searches a large database to find the most relevant information based on the customer’s query;
  • Response Generation: utilizes the retrieved information to generate a precise and contextually relevant response.

Source: Knowledge Graphs & LLMs – Multi-Hop Question Answering, Neo4j, 2023

This dual approach allows for responses that are not only accurate but also highly personalized, drawing on the latest available information. Unlike traditional models that generate answers based solely on pre-trained data, RAG chatbots dynamically retrieve information from a vast repository of knowledge sources in real-time. This ensures that the answers are not only contextually appropriate but also up-to-date, greatly reducing the likelihood of misinformation and enhancing customer trust.

There are several advantages to mention:

  • Efficiency of chatbots: by swiftly retrieving and synthesizing information, these chatbots can manage a higher volume of inquiries and more complex queries without compromising the quality of service;
  • Contextual Understanding: by analyzing the context and nuances of each request, these chatbots can tailor their responses to meet individual customer needs more effectively based on the company knowledge base;
  • Continuous Learning and Adaptation: as they constantly update their knowledge base with the latest information retrieved during their interactions, they can adapt to the clients’ most frequent queries.

Knowledge Graph (KG): making Generative AI Explainable

A knowledge graph is essentially a vast database that stores information in a structured format, using relationships between entities (like people, places, and things) to organize data. Imagine it as a network of interconnected facts where each point (node) represents a piece of information, and the lines (edges) between them describe the relationship between these pieces of information. This structure helps machines understand and use real-world information efficiently, much like a map that helps us navigate through cities by showing how different places are connected.

Source: Enhancing Interaction between Language Models and Graph Databases via a Semantic Layer, Towards Data Science, 2024

When combined with LLMs, knowledge graphs serve as a foundation for enhancing the intelligence and accuracy of these models:

  • Guide for understanding: KG provide a well-organized layer of facts and relationships. When LLMs have access to this structured data, they can better understand the context and nuances of user queries, leading to more accurate and relevant responses;
  • Better Knowledge Retrieval: KG can be a comprehensive index that the LLM can consult to fetch precise information or fill in the gaps in its knowledge. This is especially useful for answering specific questions or when detailed, factual information is required;
  • Explainability: by grounding the LLM’s responses in verified information stored in the KG, businesses can enhance the reliability of the answers provided. This is crucial for applications where accuracy is paramount, such as in financial forecasting, medical advice, or legal assistance.

How to start?

The adoption of generative AI, and specifically RAG technology, in customer care offers a compelling value proposition for businesses. The technology is evolving quickly, allowing generative chatbots to increase their accuracy in responses even for B2C applications and in contexts where adherence to sequences and procedures is critical, such as suggesting recipes from specialists and providing technical instructions to the end customer.

However, successful implementation requires careful consideration of the business’s specific needs, and the potential risks involved. The role of business managers is fundamental in order identify the scope of the use cases and the proper documentations to base RAG and KG upon and provide the feedback to regularly review the performance of AI systems and ensure they meet customer needs and company standards all over time.


  1. Osservatorio Artificial Intelligence, Risultati di Ricerca 2023, Politecnico di Milano, 2024
  2. McKinsey, The economic potential of generative AI: The next productivity frontier (June 2023)
  3. Erik Brynjolfsson, Danielle Li & Lindsey R. Raymond, Generative AI at Work, National Bureau of Economic Research (November 2023)

Do you need more information?

Fill out the dedicated form to be contacted by one of our experts.