Background and purpose Around 30% of patients with epilepsy show drug‐resistant epilepsy (DRE). While cannabidiol has demonstrated efficacy as an adjunctive treatment in Dravet syndrome (DS), Lennox–Gastaut Syndrome (LGS), and epilepsy related to tuberous sclerosis complex (TSC), its more global effectiveness in adult patients with DRE apart from these three specific contexts needs to be clarified. Methods We conducted a retrospective study at the epilepsy unit of Pitié Salpêtrière Hospital. Patients initiating pharmaceutical cannabidiol treatment and followed for at least 1 year were included. Patients were categorized into “authorized” (LGS, DS, or TSC) and “off‐label” groups. Cannabidiol effectiveness and tolerance were compared between groups, and characteristics of responders (patients with >50% reduction in seizure frequency) in the off‐label group were examined. Results Ninety‐one patients, followed by a median duration of 24 months, were included. A total of 35.2% of the patients were in the authorized group. No significant differences were observed in responder rates between groups (31.3% vs. 35.6%, p = 0.85) and retention rates at 1 year (75.0% vs. 74.6%, p = 0.97). Sleepiness was more commonly reported in the authorized group (50.0% vs. 22.0%, p = 0.01), with no other significant differences. Among off‐label patients ( n = 59), clobazam co‐prescription was more prevalent in responders (71.4% vs. 28.9%, p = 0.002). Conclusion Our findings suggest that cannabidiol may benefit all adult patients with DRE, particularly those already receiving clobazam. Randomized controlled trials are warranted in off‐label patients to validate these observational findings.