Business intelligence (BI) is the procedure of strategically planning and using a variety of tools and techniques to get important data insights and make wise business decisions. The Internet of Things (IoT) has emerged as the main source of big data with the most real-time data used across all business sectors in recent years. However, the expansion of business results is impacted by the integration of IoT as data sources with conventional BI systems. This is because there are now more security and privacy concerns in IoT ecosystems as a result of adversaries undertaking data inference and poisoning attacks on networked IoT devices via the open communication medium Internet. This study proposes an integrated architecture for strengthening security and privacy in IoT-based BI applications, which is inspired by the discussion above. The suggested structure contains two engines: an intrusion detection engine and a two-level privacy engine. Due to adversaries conducting data inference and poisoning attacks on networked IoT devices over the open communication medium Internet, there are now additional security and privacy risks in IoT ecosystems. Based on the discussion above, this study suggests an integrated architecture for enhancing security and privacy in IoT-based BI applications. The two different engines are designed namely two-level privacy and intrusion detection engine. The experimental outcomes utilising the real IoT-based dataset ToN-IoT show that the suggested strategy outperforms previous peer privacy-preserving machine learning algorithms in terms of detection rate, accuracy, F1 score, and precision.