The advent of new time domain surveys and the imminent increase in astronomical data expose the shortcomings in traditional time series analysis (such as power spectra analysis) in characterising the abundantly varied, complex and stochastic light curves of Active Galactic Nuclei (AGN). Recent applications of novel methods from non-linear dynamics have shown promise in characterising higher modes of variability and time-scales in AGN. Recurrence analysis in particular can provide complementary information about characteristic time-scales revealed by other methods, as well as probe the nature of the underlying physics in these objects. Recurrence analysis was developed to study the recurrences of dynamical trajectories in phase space, which can be constructed from one-dimensional time series such as light curves. We apply the methods of recurrence analysis to two optical light curves of Kepler-monitored AGN. We confirm the detection and period of an optical quasi-periodic oscillation in one AGN, and confirm multiple other time-scales recovered from other methods ranging from 5 days to 60 days in both objects. We detect regions in the light curves that deviate from regularity, provide evidence of determinism and non-linearity in the mechanisms underlying one light curve (KIC 9650712), and determine a linear stochastic process recovers the dominant variability in the other light curve (Zwicky 229--015). We discuss possible underlying processes driving the dynamics of the light curves and their diverse classes of variability.