Seasoned startup experience meets cutting-edge applied data science
The founder team consists of Alwin and Mel. Alwin has a PhD in applied mathematics and owns a business that develops Machine Learning projects in the field of Marketing, Pricing and IoT. Mel is an experienced founder with broad experience in Business intelligence.
A joint belief
The cemetery of unused dashboards is vast and quiet. The main reason for this, is that building useful tables, charts, maps or some other way of data visualiszation, is difficult and an iterative process.
The conventional process looks like this: the manager has a hunch, a gut feeling of what he needs. Really knowing what, is tricky, describing it virtually impossible. Also, he has no time to think it through, which is an issue because the devil is always in the detail.
The one building the solution, is left assuming and guessing pretty much everything. The first version is always not what the manager needs and after a few rounds, either the manager looses interest and accepts the last version, which is never to be used again. Or the builder looses motivation and stalls the whole thing untill the manager forgets.
Today there is a way to change this: with the use of Large Language Models, any Natural Language question can be answered in real-time, in a variety of formats. This helps the manager to iterate by himself in a matter of minutes instead of weeks. It also giver BI professionals super human powers when it comes to writing SQL or python or applying Machine Learning models
Plug & play!!!
We are far from the only ones with this idea. However, we approach the challenge a bit different than our competitors do.
The main challenge lies in the way structured databases are managed. Even the best LLM will not know what for example is in the column named ‚rev_nl_rep‘. As a result, most solutions are pretty far from plug&play: the model needs training or the databases need cleaning up: all in all, not a modern user experience and not feasible for companies that do not maintain a fit data team.
Our approach is to focus on applied and widely spread use cases, train the model on the db’s in play and reach high plug&play accuracy.
Stay tuned to see which cases are coming up.