Both skill-biased and routine-biased technological changes risk disrupting employment in Australia, particularly through persistent effects after an economic downturn. Soft skills are considered valuable for employees to reduce unemployment risk from these technological biases, as these skills contribute to employment in skilled and non-routine jobs that are difficult to automate. We investigate how soft skills affected the risk of unemployment from the Global Financial Crisis (GFC) using the Household, Income and Labour Dynamics in Australia (HILDA) longitudinal dataset to understand whether these skills could reduce unemployment risk and similar negative employment outcomes for workers during economic disruptions, including the following years during the recovery. We find that the soft skill measures of social capital and low task repetitiveness are associated with lower unemployment, overskilling, and underutilisation risk. The association between social capital and underemployment also strengthened after the downturn. This did not begin immediately after the GFC but instead from 2013 onwards, after the end of the mining boom that had supported Australia during the GFC.