Today, the Integration Center is working very well in describing its application networks and building integrations, and it has many different parties that need data. However, one department in the company needs data more than anyone else - the BI department.
The Business Intelligence department might sit with a data lake and a BI function. Many are still working with analytical BI, where one might look at the last week's data, the last month's data, and, in the best case, the last day's data. Very few are working with operational BI yet, meaning they don't understand what is happening in the reports. How do we act in our business right now to meet certain needs?
The integration centers play an extremely important role here. The BI team needs more data than any other team, and they need current data. Therefore, it is necessary to supplement batch data loading with streaming data. New solutions are needed, for example, based on Kafka instead of using ETL tools.
What is particularly interesting now is that the BI team is not only working with business intelligence; they have started working with artificial intelligence, or AI. An extremely important aspect of this is the availability of data. You can only train a good model if you have data to train on. The companies that have not yet started collecting data in their data lakes will not have any data to train on. If we see the BI development continue as quickly as it is now, including the specific area of AI, there will be a watershed within maybe a year. The companies that have data to apply their AI on and those that do NOT have data to apply their AI on. And this will then be decided by not how good a BI product you have but how good an integration product you have. How much data have you transported from your application network to your BI team?
What is required of the BI team for this to happen?
BI and integration managers need to talk more to each other. This is about today's challenge of achieving operational BI and tomorrow's challenge of getting data to train on for the coming AI function.
Otherwise, there will be no cognitive systems in the company.
No data - no cognition.