Time Varying Overdispersed Poisson Process

Is this a reasonable assumption to change the model structure like this? We know that the data is an overdispersed poisson model so I don’t think this makes much sense