We analyze a unique hybrid robo-advisory setting where portfolio management is automated, but clients are quasi-randomly assigned to human financial advisors who deliver all remaining wealth management services, including behavioral coaching and relationship support. We show that advisors of different types cause variation in client retention, especially during market downturns. To interpret these effects, we estimate a structural model in which advisors influence both investor learning about the automated solution and the perceived utility of remaining enrolled, independent of performance. Effective human advisors generate significant surplus for both clients and the firm through these informational and behavioral channels. learn more on SSRN