Define your stress periods and time steps. It is best to experiment with different time steps and compare your simulated heads to the observed values for each of the time steps you tested to determine the best time step for your model. Use the head value from your steady state as the initial head. Input your climatological data. Define your hydraulic conductivities and boundary conditions. Then run. You can check: Getting Started with MODFLOW by Richard B. Winston (2024) and Applied Groundwater Modeling by Anderson, Woessner and Hunt (2015) for a good understanding of how to set up transient models.
You can also define your first stress period as a steady state and subsequent stress periods as transient. This will reduce instabilities in your model.