At we have developed a one-stop tool linking into all systems and knowledge databases to provide employees immediate access to what they are looking for from one place.

We’ve all experienced the problem. Looking for a specific answer on your company intranet, or trying to remember the subject line of that one email you received a month back. Harvard Business Review and McKinsey found that 20-25% of a work day is spent on looking for information that we need, to get started on our tasks at work. This adds up to nearly two months a year. Our mission is to put an end to this ineffective search and simplify the whole process. Management   from right: Mogens Elsberg (Chairman of the Board), Suzanne Lauritzen (CEO), Ole Winther (CTO & Board Member), Claus Hansen (Board Member) Management

from right: Mogens Elsberg (Chairman of the Board), Suzanne Lauritzen (CEO), Ole Winther (CTO & Board Member), Claus Hansen (Board Member)

Though every system has a search function, it usually only understands keywords. We knew that in order to make search more effective we needed to do something different. Using the newest research in Deep Learning (DL) and Natural Language Processing (NLP) we came up with a way to make all knowledge searchable using natural language through a single search bar (or API integration). By linking into all back-end systems via our APIs we can process all the data and maintaining the access management rules, giving each employee access to the information they are looking for from one place.

“The problem we recognized is that companies usually use a lot of different systems and databases at the same time. Each of the databases serves a narrow purpose, but employees have a hard time figuring out where to look for the information they need. This is why in order to simplify search it was essential that there is only one place employees need to look” - CEO, Suzanne Lauritzen 

Suzanne further explains: 

“We think of the current state of the company knowledge landscape like the internet before Google. Before google search, it was impossible to find anything online. But thanks to Google’s technology all information online is  searchable and useful. We will do for the business world what Google did for the internet. With raffle, all company knowledge - whether it be files, emails or employee information on the intranet will be easily accessible and searchable, and most importantly useful. So far we haven’t met a single company that didn’t have a knowledge overload problem”.

Essentially any business in the world will benefit from the technology. Processing language has always been of great interest for data scientist. This is because language is a very distinctive part of human beings and their ability to express themselves. However, Computation with language is more complex than one might think, it isn’t just a data exhaust that needs to be processed. Essentially each word counts. The field of Deep Learning and Natural Language Processing has been preoccupied with representing words and their meaning in a way that computers can understand. This is where our unique connections to the academic environment come into play. Thanks to Ole Winther, professor at DTU and co-founder and CTO at raffle, we have a direct link to all the newest research, tools, concepts and ideas. As such our team of AI experts is working on the forefront of AI development, using state of the art models in creating a product, which is industry and language agnostic. 


Ole Winther explains“ there is a lot of talk about AI, and a lot of research is coming out. Our team of highly skilled experts takes the newest developments and we test and combine what works for us. For example, in February 2019, Facebook research came up with a new way to translate text into vector representations. The model was open source, and we took it, tweaked it, combined it with our own models, and now as a result raffle understands 93 languages.”

Besides the models that are continuously being improved, raffle is sitting on a dynamic infrastructure, leveraging microservices architecture facilitated by Kubernetes. This, alongside Microsoft Azure’s secure cloud, allows us to continuously improve the product  and add new features. “ Our technology is future proof, we make sure that the infrastructure allows us to constantly develop and improve. What many people don’t realize is that in order to have a successful AI solution, the deep learning models are only one component.. The infrastructure has to limit the amount of computational  power needed to leverage enormous amounts of data and allow new AI to be applied quickly and seamless. These are the two key areas for achieving great performance and constantly being on the forefront of technology, now and in the future.“

We live in a world where people expect everything to be just one click away, no one wants to wait for answers or attention. “We saw a gap in the market and decided to fill it ” Suzanne, CEO at raffle states. “What drives us is pushing the boundaries of what is possible with AI and seeing how our products helps thousands of people around the world become smarter and better at their work as a result of that.“