The Artificial Industrial IoT (IIoT) enabled a blend of massive different unique device varieties that can provide high quality services within decentralized computing conditions. The majority of effective AI IoT services are strongly tied to invention regulator and involve a dispersed system towards minimal latency. In case of challenging to offer constant small latency provision towards repeated requests for system services because resource arrangement built on gross traffic projections disregards. The major purpose of these actions and make it existing process of Unanticipated communication lag times are undesirable for many projects and can lead to expensive production mishaps, especially when security-related services are being provided. We propose Intellect IoT in this paper, an AI IoT provisioning system for useful services with federated learning phases-1.Industrial knowledge graph-based relation mining, globally efficient resource reservation, and federated learning-based service forecasting are the three algorithms that make up the Intellect IOT system. To improve the accuracy of service prediction, Intellect IOT combines production data with network optimisation and takes use of linkages between and within factories. The universal optimised fixed provisioning anticipated activity while taking into consideration varied in process. The analytical Intellect IOT scheme uses produce an accurate service forecast with 96.9% accuracy and improve service quality.