Introduction: This study investigates the impact of socially assistive robots and personalization in cardiac rehabilitation, in terms of physiological progress and perceptions of the patients and clinicians. Methods: We conducted a longitudinal study for 2.5 years with 43 patients. The intervention was classified into 3 conditions: control group (without a robot), a social robot and a personalized robot. Throughout the rehabilitation program (36 sessions), the initial heart rate, average heart rate during training, and heart rate recovery in the first minute. Heart rate recovery in the first minute was normalized with initial heart rate to homogenize the data: Heart rate recovery in the first minute normalized = heart rate recovery in the first minute/ initial heart rate. We apply Johansen's general formulation of the Welch-James statistic with approximate degrees of freedom on Heart rate recovery in the first minute normalized in 6 stages. Perceptions of the patients and clinicians are evaluated through surveys and interviews. Results: 26 patients completed the rehabilitation program. Heart rate recovery in the first minute normalized changed significantly in each stage of the program (TWJ (5, 305) = 14.36, p < .001). Patients assisted by social and personalized robots achieved a higher improvement in cardiovascular functioning than the control group. Both robots improved motivation and trust due to continuous monitoring and immediate feedback, facilitated prompt detection of critical conditions, and were recommended for future use by clinicians and patients. Conclusions: Our results suggest that robots improve cardiovascular functioning compared to the conventional cardiac rehabilitation program. In addition, robots have motivational benefits for patients, and increase awareness of clinicians to detect complications, as a result, they were recommended for future use by patients and clinicians.