Meet Your New Virtual Assistant – HYPERVSN x Capgemini Present ELLEM 

Generative AI has quickly transitioned from a technology reserved for those in-the-know to a tool that members of the public increasingly use for everything from improving efficiency in their jobs to deciding what to have for dinner. As consumers become used to – and sometimes disdainful of – the use of AI in tools like chatbots, it’s time for companies to find new and interactive ways of applying this technology to their customer service processes. Combining the power of generative AI with the immersive world of 3D holography is one such solution. Born from a vision to bridge the gap between human connection and technological efficiency, Capgemini’s ELLEM – a product of the HYPERVSN SmartV Digital Avatar solution – represents a groundbreaking leap forward in customer experience.

Capgemini is a global leader in consulting, dedicated to helping their clients harness the power of technology for business transformation. They discovered that 96% of CxOs have generative AI on their agenda, however, more than 90% of their projects are stuck in the proof of concept phase, effectively losing companies precious time and money. As an organisation that prides itself on enabling clients to embrace emerging technologies, creating a valuable, ready-to-go generative AI tool was a must for Capgemini. At this year’s Hannover Messe – one of the world’s leading industrial trade fairs – this new solution was unveiled: introducing ELLEM. 

The project leads – Daniela Rittmeier and Florian Erhardt, affectionately known as the mother and father of ELLEM – were inspired to create this immersive AI companion after Daniela experienced HYPERVSN’s SmartV Digital Avatar at CES 2024. She instantly knew that our life-sized digitally rendered human hologram had to be the basis of her GenAI baby! Combining our 3D holographic technology with tools from Nvidia and Mistral AI, ELLEM is a touchable generative AI communication asset with the ability to listen to you, understand your question and answer it intelligently. This AI[wo]man perfectly embodies the magic of blending futuristic technology with a human touch. Simply speak to ELLEM as you would any human (or select a pre-loaded question) and hear your answer in real-time.

Supported by an extensive in-house team, Daniela and Florian developed this trustworthy AI companion which can be deployed on-cloud, on-prem or on-board. For their launch, they built an on-prem model of ELLEM, unveiling her to an audience of over 100,000 attendees at Hannover Messe searching for innovation and inspiration. ELLEM proved herself as a potentially invaluable tool in many use cases across the entire value chain – from supporting manufacturers in their brand store rooms to providing urgent and accurate information in healthcare and urban settings. Capgemini’s ELLEM perfectly displays the capability of HYPERVSN’s SmartV Digital Avatar to help brands build lasting relationships with their customers through the power of interactivity, immersive engagement and intelligence through AI integration. 

• Interactivity gives users a personalised experience and elicits greater awareness and attention. 

• Immersive engagement delivers a powerful, surprising, and – most importantly – memorable experience for users. 

• Intelligence creates consumer satisfaction and drives efficiency in corporate applications by providing accurate information in seconds.

Thanks to ELLEM, Capgemini clients can benefit from the transformative nature of generative AI confident in the knowledge that their tool is safe, effective and secure.

But how does ELLEM actually work? There are 4 key stages that work together to drive this immersive AI virtual assistant:

• Listening: As you speak, your question is recorded with react frontend and then transcribed through OpenAI whisper model. 

• Understanding: context is delivered through advanced search functionality of an extensive database. This is achieved with a Nomic embedder and Qdrant Vector DB. 

• Processing: A large language model (LLM) processes your prompts and the instructions and information from the previous step, in this case, Mistral AI. The answer is then transcribed by Azure offline text to speech. 

• Answering: The answer is presented in an immersive way through the HYPERVSN SmartV Digital Avatar.  

“What really excited me about this project is that we decided to go on-prem with the whole solution. The curated data we use from Capgemini to answer the questions with the right context is working really well, [and] we’ve already received very positive feedback [about ELLEM] .”

– Florian Erhardt, Senior Generative AI Architect, Capgemini Germany

Here at HYPERVSN, we’re excited to see our 3D holographic technology brought to life in such an inspiring and thoughtful way. Using our SmartV Digital Avatar solution, AI companions like ELLEM have the potential to impact consumers of all backgrounds and in different contexts. The ability to modify the avatar’s appearance and voice means that our holographic companion can reflect the diversity of the world around us, giving a truly human touch to its interactions. While many are still cautious when it comes to AI technology, over half of global consumers (51.5%) describe themselves as either somewhat or very positive towards generative AI. It is our role to educate and inspire consumers by showing what is possible. Imagine a multilingual avatar stationed in airports around the world, or a new, engaging support tool for overstretched teachers. Imagine AI companions providing world-class customer service in banks, shops, hospitals, and other places where we see fewer and fewer support staff. With interest from such a wide range of industries, we can’t wait to see where this technology goes next!