DataRobot does deliver, but needs time

We are mainly using it as a tool to develop ai knowledge and try to take our first steps. For example product recommendations, delivery time prediction, and sales predictions. We are still extremely new and are not currently running a system in live production. But we are seeing the value and are making progress.

Some projects were done exclusively through the platform, but there were instances where it was beneficial to use outside resources. For example for working on data (visualization, joining data) and also for building simpler prototypes and first versions of statistical models instead of artificial intelligence. But overall the integrationist was helpful.

We were hindered especially by providing all necessary data in the necessary quality. And by other factors. We are now 6 months into the use of DataRobot and don't have a running ai in production. We started the first few weeks with planning and ideation, then weeks with data exploration and non ai solutions, and are only now building the first ai driven solution for delivery estimation.