Looking back I see I was a bit lazy just giving activities related to a guest room as examples. @george, I hope this post makes amends.
guestroom > booking
Here a company like Trivago asks its web site visitors to select a location, bed type and check in/out dates. A protocol and mechanism exists for the relevant data to be extracted directly or indirectly from each hotel website for it to be analysed and returned as a clickable display of compliant deals.
guestroom > furnishing
In this scenario, it is the hotel owner/designer/operator who is seeking information for one or perhaps hundreds of rooms. The search is for multiple items - e.g. bed, nightstand, desk, minibar, luggage stand, closet, seat, sofa - as accountable selection is based not only on price but also compatibility of sizes, delivery, distribution, maintenance, country of origin and so on. Universal protocols and mechanisms are designed for such complexity.
The first search pass finds the sites that hold the data needed so that aspect pairs can be extracted and analysed to return unique keys and values. These are marked up to exclude those not required and next pass run. Clearly there is pressure to balance accountability with time available for the re-running, but this could be significantly relieved by devising conditional criteria, i.e. more sophisticated mechanism design.
I mentioned 3D modelling in my last post because this is important where subjective judgment is part of decision making. It also enables animation to be used to explain processes and visually track progress using input from diverse sources - the delivery guy, for example.
protocols and mechanisms
Those set up for furnishing also work for the simpler booking example, the company’s role undertaken by individuals’ JSON machines, with relevant imported monitors to generate different types of display as required.