Audi revolutionizes internal documentation with RAG-based AI chatbot
The car manufacturer Audi now relies on Generative AI, which together with Retrieval-Augmented Generation (RAG) ensures dependable internal knowledge management.
#Retrieval-Augmented Generation
#Generative AI Chatbot
#Automotive Industry
,allowExpansion)
The Vision
An AI chatbot for internal knowledge management.
The Scenario
Ensuring efficient knowledge management across departments
Whether it's project documentation, risk assessment, or just a colleague's phone number: Employees need a large amount of different internal information every day. However, searching for this information is often time-consuming and not always effective.
The car manufacturer Audi wanted to counteract this problem. A chatbot was supposed to provide employees with the information they need in an uncomplicated manner in just a few seconds, thus improving productivity and the quality of work results and decisions.
The challenge
Receive correct and specific answers
Because common chatbots often use outdated technologies such as simple keyword searches, they frequently result in incorrect or very general responses. AI-based chatbots also pose risks: If not enough suitable data is available, they often create invented answers. All of this can have fatal consequences, especially within companies, and lead to inefficient processes or incorrect decisions.
The solution
A new type of AI chatbot from Storm Reply
Hallucinations and unspecific or outdated responses? Not an issue for Audi employees using the new chatbot. It's an innovative AI solution that Storm Reply's experts developed within four weeks based on the Amazon SageMaker service. If the AI chatbot cannot answer the posed question with certainty due to insufficient data, it uncovers this instead of constructing an answer. This way, employees can use the information provided without hesitation and count on the results. In addition, the chatbot does not provide general answers, but only such answers that are relevant for Audi.
)
How we did it
Taking generative AI to a new level with RAG
Retrieval-Augmented Generation (RAG) makes all of this possible. This technology, in combination with state-of-the-art prompt engineering, allows the output of the AI chatbot and the Large Language Model (LLM) that lies behind it to be selectively optimized with specific information. The model itself remains unchanged in the process. Before the chatbot gives an answer, it first retrieves all relevant entries from an external base specially defined by Audi. The car manufacturer can flexibly adapt and expand this at any time. In this way, it can guarantee not only correct answers but also answers that are always up-to-date.
This chatbot is a great example of how innovative technology can be used to solve practical use cases within a very short time-to-market. It reduces the barriers to finding information for our internal stakeholders.
Michael Pawelke
Product Owner AWS Foundation Services, Audi AG
The advantages
Uncomplicated, fast, and secure
The new internal AI chatbot gives employees at Audi uncomplicated, fast, and secure access to information. This allows the company to break down knowledge silos and realize better decisions, more efficient processes, and ultimately cost reductions. In addition, the chatbot offers the following benefits:
Security by design
The experts designed the chatbot architecture within a virtual private cloud (VPC) to ensure data security and actor privacy.
New insights
The chatbot not only reduces search time from hours to a few seconds but is also able to draw conclusions and generate unique insights from the data.
Multiple sources of information
From Confluence to Notion, from PDF to custom websites- our AI chatbot solution is able to ingest data from a variety of source and keep the knowledge base constantly updated.
The Audi Group is one of the most successful manufacturers of automobiles and motorcycles in the premium and luxury segments. The brands Audi, Bentley, Lamborghini and Ducati produce at 21 locations in 12 countries. Audi and its partners are present in more than 100 markets worldwide. In 2022, the Audi Group delivered 1.61 million Audi brand cars, 15,174 Bentley brand cars, 9,233 Lamborghini brand cars and 61,562 Ducati brand motorcycles to customers. In the 2022 fiscal year, the Audi Group generated an operating profit of €7.6 billion on revenue of €61.8 billion. More than 87,000 people worked for the Audi Group worldwide in 2022, including more than 54,000 at AUDI AG in Germany. With its attractive brands as well as new models, innovative mobility offerings and pioneering services, the Company is systematically continuing on its path to becoming a provider of sustainable, individual premium mobility.
Storm Reply is specialized in the design and implementation of innovative Cloud-based solutions and services. Through consolidated expertise in the creation and management of Infrastructure as a Service (IaaS), Software as a Service (SaaS), and Platform as a Service (PaaS) Cloud solutions, Storm Reply supports important companies in Europe and all over the world in the implementation of Cloud-based systems and applications. Storm Reply is AWS Premier Consulting Partner.