Industry 4.0 (I4.0) represents a transformative approach, integrating technology, production methods, and information and communication technology to enhance industrial value creation. A central I4.0 goal in the energy domain is improving energy efficiency to boost industrial competitiveness and profitability. Given that induction motors account for nearly two-thirds of industrial electrical energy consumption, optimizing their efficiency is crucial. Energy management systems (EMSs) need real-time data to assess motor efficiency, enabling prompt identification and replacement of inefficient motors with alternatives that have optimal efficiency class and rated power for specific applications. This paper introduces a novel non-intrusive method for estimating the load and efficiency of induction motors without disrupting their operation. To reach that goal, the proposed method optimizes the parameters of a set of relationships between output power, input power, and losses with the motor speed, minimizing the error in the estimates. It requires only input electrical power and motor speed measurements to set the model parameters and estimates the load and efficiency using either speed or input power measurements. The experimental results demonstrate that the proposed method, with a mean overall error of less than 3.5% in estimating output power and efficiency, outperforms conventional methods.