ogGPT: Otto Group launches internal, privacy-compliant tool with generative AI for all employees

Artificial intelligence (AI) will fundamentally transform the working world. AI assistants like OpenAI’s ChatGPT already offer numerous opportunities for companies and their employees to optimize workflows and increase efficiency. However, ensuring the compliant use of publicly available AI assistants within a company can be challenging due to various regulations and strict usage restrictions.

With ogGPT, the Otto Group has developed its own secure and privacy-compliant solution in a very short time and made it available to approximately 26,000 employees in the DACH region at the end of September. In the first week alone, the chatbot was used by more than 4,200 employees. A company-wide rollout is already being prepared.

The in-house AI language model was conceptualized and implemented within a few weeks by a cross-functional team of experts from the Otto Group IT and Digital & Consulting departments. Chats and data remain under the control of the Otto Group IT at all times and are not used for training the underlying AI models. To achieve this, the company relies on the Microsoft Azure OpenAI Service hosted in the EU, which allows companies to utilize the latest generative AI models from OpenAI, such as ChatGPT, GPT-4, or DALL-E, through Azure.

In addition to the chat function, employees can also upload their own documents or entire knowledge bases and interact with them through the new AI assistant, for example by asking questions about the content. The service then extracts the relevant passages and uses them to answer the questions – and this can be done in multiple languages. The function of source references is particularly valuable, allowing for easy traceability of the generated answers. Translations can also be created effortlessly.

"The potential of artificial intelligence is enormous, and we are currently only scratching the surface. No other technology will be as disruptive as AI in the coming years. I am deeply impressed by how quickly our teams integrated our own secure and privacy-compliant version of ChatGPT into our systems. This allows our colleagues to easily and safely utilize generative artificial intelligence in their daily work," explains Sebastian Klauke, Otto Group Executive Board Member for E-Commerce, Technology, Business Intelligence, and Corporate Ventures. He adds: "Of course, we adhere to our internal data protection and information security policies when using AI-generated output, which means we review it professionally in advance and appropriately label it."

Employees receive support in dealing with generative artificial intelligence (AI) through the internal Discover-AI community, a dedicated support unit that offers specific assistance and specialized online training tailored to the needs and requirements of various AI user groups within the company, including offerings for technical experts as well as beginners. Furthermore, employees have access to additional video-based learning paths on topics such as digital transformation and specifically AI through the company-wide learning ecosystem, the Otto Group Learning Experience Hub.

"This is just the beginning. We will continuously develop applications like ogGPT to offer even more exciting possibilities for all colleagues in the field of generative AI. Our goal is to constantly explore new ways of utilizing generative AI and, of course, to learn together with our customers how these technological possibilities can further enhance online retail and the customer experience," explains Alexander Birken, CEO of the Otto Group.