This paper surveys urban courier routing, pointing out the learning process of the generalized travel cost enhanced by using innovations related to the introduction of emerging information and communication technologies (ICTs, i.e., the internet of things, big data, block chain and artificial intelligence), considering a smart city. Couriers, when planning in advance or choosing the routes in real time for delivering to citizens as well as to business users (including retailers), need to consider both the driving and walking routes (i.e., from the delivery bay to the customers) to optimize their activities. A two-layer literature optimization model is recalled, and the main scientific people-centered challenges that need to be addressed under the light of emerging ICTs are identified and explored, which are the learning process of routing attributes, as well as the opportunity to book on-street delivery bays in advance or in real time. Then, after a literature review on modeling courier activities, a unitary formulation is presented that combines old and real-time network data. In addition, integration with new telematics solutions (i.e., delivery bay booking) is pointed out. Finally, discussions on innovations and cost optimization are presented.