I need a publically available one-to-one speech conversation dataset recorded over the microphone, e.g., interview scenario or enquiry or call centre call answering.
Searching the Virtual Language Observatory maintained by the European CLARIN ERIC is the way to go. Enter
conversation spoken in the search slit and narrow down the search using the facets (e.g., by chosing a language). Here is a bookmarked query on the VLO giving a lot of resources for English language.