AI is increasingly being adopted in manufacturing due to Industry 4.0 and the availability of data, enabling dynamic and connected environments that improve agility, productivity, and sustainability. Solutions like predictive analytics and assisted decision-making enhance product quality, performance, and cost efficiency. However, challenges such as real-time processing, security, integration of diverse devices, and large data volumes hinder industrial AI adoption. The "Plug & Produce" concept offers a solution by enabling seamless integration of new components into production systems without manual efforts. Current AI pipelines are often disconnected from shop floors, relying on manual data input and lacking automatic output integration. Achieving seamless AI integration requires an IIoT connectivity framework spanning physical, transport, and data layers. The lack of clarity in linking AI applications with machines and the diversity of available technologies complicate decision-making. The ICNAP study addresses these challenges by providing an overview of technologies, establishing key selection criteria, and developing a structured integration procedure. These guidelines aim to simplify technology research and help businesses integrate AI more efficiently. This approach supports faster implementation and greater adaptability in manufacturing environments.