In 258 US adults with problematic substance use (202 analysed), a smartphone relational-agent app did not reduce past-month substance use occasions at 8 weeks vs emailed psychoeducation (beta -0.675, p=0.741); both groups improved.
RCTn = 258
Prochaska et al. A relational agent for treating substance use in adults: A randomized controlled trial with a psychoeducational comparator