The complexity of dynamic environments in which companies work requires their managers to take rapid and relevant decisions. A business activity monitoring application should be able to implement complex bi-temporal queries on real-time data and historical data in order to detect business trends and anomalies. However, accessing both real-time and historical data is costly and can hardly meet the application’s requirements of fast responses for analysts.In this paper, we propose Decision Insight, a platform developed by a French software editor, that solves this problem. It consists in redefining such a complex bi-temporal query into: 1) a set of continuous queries in charge of handling real-time data (whose results are materialized) and 2) a query that accesses both historical and materialized results of the previous continuous queries. Thus, Decision Insight can provide analysts with timely answers through a convenient GUI.Decision Insight is based on a column-store bi-temporal DBMS that handles these two types of data and implements a simple and efficient bi-temporal query optimization technique. We demonstrate the interest of our approach using an adapted version of TPCBiH, a bi-temporal extension of the TPC-H benchmark. Extensive experiments have been conducted, pointing out the interest of Decision Insight for delivering timely information based on historical and real-time data.