In recent decades, broadcast archives have opened up their collections with automatic speech recognition (ASR). For archives, ASR is predominantly used as a form of metadata enrichment to increase the chances of retrieval: ASR translates spoken dialogues into text and enriches the metadata, therefore increasing the chances of discovery (e.g. Ordelman and van Hessen 2018). Within scholarly research, the speech transcripts are most often used as a convenient source for discourse analysis, as they provide a representation of what has been said in the broadcast. In this research, we will move beyond the use of ASR-data as mere transcripts. By approaching ASR-data ‘with care’, and thus maintaining and respecting its characteristics, we will shift perspectives to the methodological opportunities and challenges of this data.