We employ the Planck 2013 CMB temperature anisotropy and lensing data, and baryon acoustic oscillation (BAO) data to constrain a phenomenological $w$CDM model, where dark matter and dark energy interact. We assume time-dependent equation of state parameter for dark energy, and treat dark matter and dark energy as fluids whose energy-exchange rate is proportional to the dark-matter density. The CMB data alone leave a strong degeneracy between the interaction rate and the physical CDM density parameter today, $\omega_c$, allowing a large interaction rate $|\Gamma| \sim H_0$. However, as has been known for a while, the BAO data break this degeneracy. Moreover, we exploit the CMB lensing potential likelihood, which probes the matter perturbations at redshift $z \sim 2$ and is very sensitive to the growth of structure, and hence one of the tools for discerning between the $\Lambda$CDM model and its alternatives. However, we find that in the non-phantom models ($w_{\mathrm{de}}>-1$), the constraints remain unchanged by the inclusion of the lensing data and consistent with zero interaction, $-0.14 < \Gamma/H_0 < 0.02$ at 95\% CL. On the contrary, in the phantom models ($w_{\mathrm{de}}<-1$), energy transfer from dark energy to dark matter is moderately favoured over the non-interacting model; $-0.57 < \Gamma/H_0 < -0.10$ at 95\% CL with CMB+BAO, while addition of the lensing data shifts this to $-0.46 < \Gamma/H_0 < -0.01$.