BACKGROUND: Adherence to therapies and metabolic control among patients with type 2 diabetes mellitus (T2DM) remain challenging. The use of new technologies, such as telemedicine, digitalized systems, and social networks, could improve self-management and disease control. OBJECTIVE: We evaluated the efficacy of a digital educational intervention for patients with T2DM, expressed as changes in glycated hemoglobin (HbA) and body composition and evaluation of the response using validated questionnaires of satisfaction with health care professionals (Instrument for Evaluation of the Experience of Chronic Patients), Diabetes Knowledge Scale (ECODI), and adherence to treatment over 6 months of follow-up (Morisky, Green, Levine Medication Assessment Questionnaire). METHODS: This multicenter, randomized, prospective study included adults with T2DM with poor metabolic control who started treatment with glucagon-like peptide-1 receptor agonists. Patients were randomized to digital intervention or usual care. The intervention group received education through social networks and digital tools in a structured program of healthy lifestyle changes. This was provided by a "Digital Coach" for weekly and on-demand advice and individualized support. Baseline and follow-up demographic, clinical parameter, adherence, and quality of life data were collected. RESULTS: We included 85 patients (control: n=41; intervention: n=44). Both groups were matched regarding demographics, physical examination, insulin, and biochemical parameters. We observed a reduction in body weight (intervention: -8.7, SD 6.1 kg vs control: -4.9, SD 5.0 kg; t=-3.13; P=.002), BMI (intervention: -3.0, SD 2.1 kg/m vs control: -1.8, SD 1.8 kg/m; t=-2.82; P=.006), and fast mass in both groups but greater in the intervention group. There were greater reductions in fasting plasma glucose (intervention: 122.6, SD 81.5 mg/dL vs control: 70.5, SD 72.9 mg/dL; t=3.10; P=.004) and HbA (intervention: 3.7%, SD 1.9% vs control: 2.6%, SD 2.1%; t=2.54; P=.006) in the intervention group. Although there was no significant change in the Spanish version of the Diabetes Quality of Life Questionnaire (EsDQOL) satisfaction score in the control group after 6 months of follow-up (0.7, SD 19.8), there was a marked reduction in EsDQOL satisfaction score in the intervention group (-13.7, SD 23.1; t=-3.08; P=.02). According to the ECODI scale, knowledge about diabetes increased more in the intervention group (intervention: 0.3, SD 1.8 vs control: 1.5, SD 1.5; t=-3.33; P=.001). Although the medication adherence score worsened in the control group after 6 months, it significantly improved with the intervention (control: -8% vs intervention: 13.8%; χ=0.35; P=.01). Patients' health care experiences improved with the intervention but not with the control. CONCLUSIONS: The digital educational intervention was effective at improving glycemic control, body composition, adherence, and patient satisfaction compared with usual care in patients with T2DM. The implementation of digital tools and social media could highly improve the multidisciplinary approach to the management of this population. TRIAL REGISTRATION: ClinicalTrials.gov NCT06850129; https://clinicaltrials.gov/study/NCT06850129.