Data warehouse

A data warehouse pulls data from your CRM, accounting, ERP and other systems, puts it into one fixed structure and keeps the history. That makes it the classic heart of business intelligence. Whether your data foundation for AI needs one depends on what the AI has to do: analyse years of data, or work inside a running process.

Loading usually happens periodically, for example every night, through a process that extracts, transforms and loads the data (ETL or ELT). That makes a data warehouse strong at trends and comparisons across years and systems, and less suited to acting directly within a running process.

For AI that helps with day-to-day work, current data from the source systems often matters more. You get that with system integration. A data warehouse is added when the analysis question calls for it, and only delivers reliable figures once data quality at the source is in order. How AI then works across your connected systems is covered under the AI Operating System.