Business Intelligence 1.0 is defined as specialist, small-audience BI tools such as Discover for OLAP, ProClarity and Cognos Powerplay, datasets that are largely database-based and with latencies of a day or a week, data sources confined to formal, dimension-style data warehouses, the actual analysis of the data being a separate task done away from line of business applications, and expensive, proprietary toolsets brought together from a number of nice, specialist tool vendors.

Business Intelligence 2.0, in contrast, is defined by large, non-specialist audiences, BI delivered directly into line of business applications rather than requiring a context shift to a separate OLAP tool, datasets made up of event and service-based data sources as well as traditional data warehouses, analysis and decisions being taken automatically by the BI infrastructure when the event happens, rather than manually, after-the-event by a team of analysts, real-time data in order support better operational decisions, tools and dashboards as easy to use as Google rather than ones that require a week’s training to use, and loosely-coupled, “mashed up” decision architectures brought together quickly to react to a business opportunity.

So, to reduce this list down to it’s essence, Business Intelligence 2.0 seems to come down to:

• Event and service-based data sources (i.e. from middleware) as well as warehouse and operational data from the database
• BI embedded in business processes and applications
• Business processes callable from BI
• Flexible, loosely-coupled and quick-to-develop data architectures
• Real-time data
• A much wider audience (implications on scalability, complexity of user interfaces)
• Automated decisions