James Kobielus of Forrester posted on BI’s New Frontiers In 2008 And Beyond. While James made some interesting points I think that he missed perhaps the most interesting one – whether what we call “BI” today will be able to evolve into what we need in the future or whether these new frontiers will require a whole different mindset.
In particular, the widespread focus on “operational BI” and James’ comment about BI’s role in Complex Event Processing and SOA, raise an important issue. Can BI tools, focused on how to present information to people so that they can gain insight ever work in an environment where the insight is required by an automated system? Think about it – BI tools process data to present information to people so that the people can make better decisions. Yet Complex Event Processing and SOA require that systems use the available information to make better decisions. Visualization, reporting, OLAP, Excel, pivot tables and all the rest of the paraphernalia of BI are useless as systems, services, cannot use them. To apply data in these real-time operational systems you need to be able to turn that data into something both useful and executable – a predictive analytic model. This is why, although enterprise decision management builds on BI to some extent, EDM is a replacement for what many call “operational BI”. If you want to use your data to make better decisions in your operational systems, you don’t need BI, you need EDM.