Objective Interpersonal psychotherapy (IPT) has demonstrated efficacy for the treatment of depression, yet little is known about its therapeutic mechanisms. As a specific treatment, IPT has been shown to directly reduce depressive symptoms, although it is unclear whether these reductions occur via interpersonal changes. Within IPT, the potential role of the working alliance, a common factor, as a predictor of depression and interpersonal changes is also unclear. Method Participants were 147 depressed persons living with HIV in rural communities of 28 U.S. states enrolled in a randomized clinical trial. Seventy‐five patients received up to 9 sessions of telephone‐administered IPT (tele‐IPT) plus standard care and 72 patients received standard care only. Two models were tested; one included treatment condition (tele‐IPT vs. control) and another included the working alliance as independent variables. Results The first model found an indirect effect whereby tele‐IPT reduced depression via decreased social avoidance. There was a direct effect between tele‐IPT and reduced depression. In the second model, the working alliance influenced depressive symptom relief via reductions in social avoidance. Both goal and task working alliance subscales were indirectly associated with reductions in depressive symptoms, also through reductions in social avoidance. There were no direct effects involving the working alliance. Tele‐IPT's influence on depressive symptom reduction was primarily through a direct effect, whereas the influence of working alliance depression was almost entirely via an indirect effect through interpersonal problems. Conclusion Study findings have implications for IPT when intervening with depressed rural people living with HIV/AIDS over the telephone.