Can training and mentorship expand the economic impact of cash transfer programs, or would such extensions waste resources that recipients could allocate more impactfully by themselves? Over the course of two years, a Ugandan nonprofit organization implemented alternative poverty alleviation approaches in a randomized manner. These included an integrated graduation-style program involving cash transfers as well as extensive training and mentorship; a slightly simplified variant excluding training on savings group formation; and a radically simplified approach that monetized all intangibles and delivered cash only. Light-touch behavioral extensions involving goal-setting and plan-making were also implemented with some cash transfer recipients. We find that simplifying the integrated program tended to erode its impact.