In India, two-thirds of the population relies on the agricultural sector for their livelihoods. Leveraging Artificial Intelligence (AI) and Internet of Things (IoT) in agriculture brings in a new era of connectedness through computer-based systems. Irrigation is the sole of agriculture, but it is becoming a concern not only in India also across different parts of the world, due to changing climatic conditions and its impacts on soil. To address this challenge, this paper proposes an AI-IoT-based framework, which consists of three phases namely, (i) Plant health classification, (ii) Irrigation Recommendation/Control based on Plant Health and Soil parameters and (iii) Mobile application-based tracking for yield monitoring and plant management. For classifying the crops as healthy or unhealthy, Convolutional Neural Network (CNN) model is trained with plant-disease dataset. Using 7-in-1 sensor, the soil parameters namely N, P, K, pH, soil moisture, temperature, and humidity are collected. Long Short-Term Memory Model (LSTM) is trained using the plant health status and soil parameters thereby predicting the level of irrigation. Raspberry pi 3 b + is employed to pre-train CNN and LSTM models, for irrigation recommendation of the indigenous crops that incorporate climate change adaptation. Also, this paper offers farmers crucial insights through a user-friendly mobile app named Plantio. From the results obtained, it is proved that the proposed approach enables the farmers to monitor the plant health status and environmental parameters remotely and control the irrigation schedule conveniently. The CNN model showed 87% classification accuracy, and the LSTM model demonstrated acceptable validation accuracy and validation loss, making this framework robust and efficient.