Xayn launches new enterprise AI search that’s energy and cost efficient

BERLIN. May 2023 – The research-based AI company Xayn presents today Xaynia, the enterprise semantic search and recommendation AI. It combines accuracy, energy- and cost-efficiency with privacy. Business clients across different industries can easily implement the AI plug-in via API to provide their customers with a smooth and individually personalised user experience. Xaynia moves with its Natural Language Processing (NLP) technology from matching to understanding its users’ search queries and their context and delivers highly accurate results tailored to individual preferences.  

Energy- and cost-efficiency through Small Language Models  

“While Large Language Models like ChatGPT have created quite a buzz they also consume huge amounts of energy – at the cost of the environment. This is not only a problem for the climate, but it also increases the costs for business clients immensely, especially when used at scale. Additionally, we don’t know what happens with all of the personal data," says Leif-Nissen Lundbæk, PhD (CEO & Co-Founder of Xayn). "Our suite of multilingual NLP models Xaynia solves these problems for search and recommendations by moving from Large to Small Language Models that are even so efficient that they can run on a mobile phone instead of the latest supercomputer.”  

Keeping the accuracy of Large Language Models, Xayn compressed its AI to maximise the energy efficiency of Xaynia. This way, Xaynia can process Big Data such as millions of documents in large databases while minimising energy consumption and costs at the same time. In a recently published paper, Xayn tested its own model extensively on energy efficiency comparing it to other popular transformer models which are used to understand users’ search intentions and indexed contents. In these settings, Xaynia was up to 20 times more energy efficient.  

Xayn is validating its technology with different media companies such as the German public-broadcast TV station ZDF. The media institution is aiming to improve their media library ZDFmediathek by personalising the recommendations for their users in a more privacy-protecting and energy-efficient way. 

Diverse applications across different industries 

With personalisation proven to increase revenues and achieve higher retention rates, Xayn’s versatile semantic search and recommendations plug-in addresses the needs of companies with large outward-facing knowledge bases from sectors such as media, e-commerce, or travel. Catering to the users who don't actively state their preferences, personalised search results and recommendations ensure content relevance, enhancing the overall user experience. Xaynia can be applied to websites, apps, push notifications, and newsletters, creating a cohesive and engaging experience for users across various platforms. 

European standards in transparency, privacy and sustainability 

The Berlin-based company originated in Oxford and is committed to the European values of transparency, privacy, and sustainability and is pushing for Europe’s technological sovereignty. Therefore, it is developing Xaynia as a privacy-protecting open-source solution. On the paid plan, Xayn’s business clients will benefit from an easy-to-integrate API solution. It enables them to utilise Xaynia and Xayn’s highly efficient infrastructure without the need for machine learning expertise or maintaining a complex content discovery system themselves. 

For more information go to www.xayn.com