In Wizdee, we have always looked at the world of Business Intelligence solutions with the sense that these could be more simple. We are mainly from the Artificial Intelligence area, but with a good background (and several Industry projects) in Business Intelligence, and it was always a big pain the amount of work and complexity that the BI projects demanded from us.
The enormous amount of energy that needs to be spent just modelling the data warehouse/mart (the famous star model), the need to predefine all the reports and analysis that the users can do, and other complex tasks that need to be done before the system goes into the hands of the Business users.
Then the users start using the system and they want more … a new report, new dashboards, new listings, well you know it. You give the business users a tool that enables them to take a peek at their data and they want more, always more. The problem is that these projects are endless and take a lot of time, leaving the users with an aversion to explore the data freely. These projects always left us with the bitter sense of coming short to the user’s expectations, which is something that we have difficulty in living with.
We then started to ask ourselves: what if there was a way to create a solution flexible enough, so that the business users are able to create their own reports and dashboards? what if there were a solution intelligent enough to discover associations in the data enabling the user (with no technical skills) to explore the data and discover valuable insights? We started asking ‘why not’ instead of ‘why’ (see the Thomas Korte talk https://www.youtube.com/watch?v=YVLaEvjtC6A). We were the guys from the AI side of the story, so we carried no dogmas from traditional BI.
At the same time and back in 2009, from the several projects that our team was involved in, one of them (and more in our domain of expertise) was a semantic search engine. A search engine capable of understanding the user’s queries and decide what would be the right answer. This was a solution for structured information, but only for searching information.
One day, in a conversation with one of our technological partners, he started to provide feedback about our ideas for the semantic search engine. This conversation was very inspiring and lead us to see things in a new perspective and ask: “Why not using a semantic search engine to do BI?”, “Why is not the user able to do a simple natural language query to have the analysis that s/he wants?”. This made perfect sense for us, more: it felt right! We could easily understood this feeling, if we associate this vision to the idea that we have on user experience: it should be terribly simple. One of the worst things that we felt about most of the information systems that we know and use, is that they are made for developers/technicians and not for the common user. So they have lots of options, menus, commands, an overload of visual information that makes the user feel lost.
After our epiphany, we started trying to operationalize it. Our vision was clear: empower the business user with a solution where s/he could type “count customers by state” and get an answer to that query. Simple. Fast. Effective. To implement this vision we turned to our semantic search engine, which could already deal with natural language, we ‘just’ needed to add more expressiveness to be able to deal with analysis and operations.
After one year of development, we finally got to our first prototype. We had to restructure and improve our engine, but we found out that the user interaction would have to be different. At the image of the search engine, the user experience had to be simple, intuitive, and above all, effective. In the fall of 2013 we started looking for funding to take our idea to the next level, and in last May, we closed our seed round, with 1M USD from Portugal Ventures and Novabase Capital, which share our vision and back it up.
We want to provide the users with a solution that is truly a self-service solution for the business user. A solution where the user does not need to be a technician or a power user. A solution where an everyday user can just search and analyze information as if s/he was using ‘the famous search engine’, just type the question in natural language and gets her/his answer. That simple!