MediaMarktSaturn ignites the GenAI turbo
Generative AI is on everyone's lips, in the news, on social media and now also freely accessible to everyone on the web. Some see it as a revolution in the way we will work, live and interact with each other. Others are skeptical and still see many unanswered questions.
That's why we started implementing artificial intelligence some time ago. For example, with our chat bot Emmi on our websites, which we have trained so that it can answer simple customer inquiries about opening hours or product information. Another field is the creation of product texts using artificial intelligence. By entering the available information, pre-formulated texts with descriptions and product data can be created for the corresponding product detail page on our website.
In an initial idea and implementation phase, the company's internal GenAI project group developed initial pilots for previously selected use cases with colleagues from the business and from MMS Technology as well as other partners consisting of Accenture, Accesa, Google, Microsoft and Parloa.
We always see the use of Generative AI in interaction with people - customers and, above all, our employees. Generative AI will support our employees in making better and faster decisions. It will not replace employees.
"Work-AI-Balance"? The useful helper for everyday working life.
We want to support our employees in expanding their knowledge of generative AI and building up new knowledge. We offer training courses ranging from initial onboarding to in-depth specialist knowledge of Generative AI.
We also offer a curated library that provides quick access to inspiration and knowledge. And we provide our employees with an internal platform with corresponding generative AI applications that they can use for their daily work in a secure, corporate-approved environment. As a first step, we are implementing the AI Enabler Platform and the GenAI Sandbox, followed by the "42", a Digital Workplace assistant.
GenAI Sandbox
The aim is to provide employees with an AI-based knowledge management tool that enables them to find the information they need more quickly and efficiently. For the pilot project, a chat assistant will collect and process relevant information from Sharepoint (our intranet) and other connected information sources in the company and output it in a way that is tailored to the employees' questions.
"42" by Digital Workplace
The idea behind "42" is to act as a central entry point to all services offered by Digital Workplace. Powered by GPT-4, it can take actions and answer questions about internal knowledge coming from Sharepoint (our Intranet), or our IT ticketing system knowledge base, as well as general knowledge, and by connecting to several internal systems it can also book calendar appointments, holidays, open tickets on the user’s behalf and much more.
AI Enabler Platform
The scalable AI Enabler Platform based on the sandbox is the basis for implementing various LLMs as prefabricated models with data protection-compliant requirements and corresponding access management in the company system. This ensures that new AI projects are implemented quickly, cost-effectively and with the necessary requirements.
How can I help you?
We also see great potential in contact with our customers to make certain processes even more efficient and solution-oriented or even to create new experiences. Longer waiting times for a free call center employee at peak times are now a thing of the past, as are language barriers. Old-fashioned operating instructions in handout format come to life and are interactively controlled by the person seeking help. Generative AI has the potential to improve many tasks, activities and processes for customers, to make them simpler and more convenient and, on the other hand, to optimize business processes or relieve our employees of smaller standard tasks.
Customer Care Bot (chat and voice bot)
We combine our existing bot "Emmi" with an LLM (Large Language Model) and enable the bot to interact more fluidly and smoothly with customers. The aim is to integrate a generative AI-based customer care bot as a chat and voice bot. The bots should be able to answer FAQ-related questions.
SEO-optimized Content Creation
AI can be used to deliver automatically generated content that contains the tone of voice of industry partners and is optimized for keywords. Content must be adaptable in terms of tone of voice and emotional appeal. The first pilot will be able to process PDP content, headlines and keywords.
Interactive Instruction Manual
With the interactive instructions, potential problems and sources of error can be identified and rectified at an early stage. Customers can interact in their own language by describing their problem to the bot. The bot responds in the customer's language. Additional potential: Digitalization allows problems and specific questions to be collected and resolved in future before they occur. Additional services can also be offered to prevent such obstacles from arising in the future.
Sales Assistant Bot
The Sales Assistant Bot can provide product advice and recommendations in our online store or in the brick-and-mortar store. It is implemented as a chat or voice bot in combination with a human-like avatar. An initial pilot is expected to be available to employees at our head office for initial tests.
(Live) Translation
The AI can translate content into different languages, adapt it to the country-specific use case and keywords and modify it to avoid duplicate content. Translation is primarily used for customer-oriented content, documents and services. Another area of application is the real-time translation of customer inquiries in call centers, which is expected to be used in an initial test phase in the Netherlands at the beginning of 2024. Both the customer and the employees in the call center will receive the conversation translated by the AI into their respective national language, sometimes even with regional accents.
The Future is... GenAI?!
The first phase of implementing the GenAI projects has started. The applications are being continuously optimized and initial successes and added value are quickly becoming apparent. But this was just the starting signal for a whole series of further Generative AI projects. In the first phase, we were already able to collect over 100 ideas that have the potential to be implemented as a separate project.
And there are more and more every day: from the wealth of ideas of around 50,000 internationally active colleagues at MediaMarktSaturn, from the exchange with partners and the further development of existing applications, we will work every day to implement and establish generative AI applications at MediaMarktSaturn. Let's Go!
External publication on artificial intelligence at MediaMarktSaturn
