In AI work, data covers far more than rows in a spreadsheet. It’s emails, transcripts, client notes, PDFs, call recordings, the messy trail of how work actually happens. If information is trapped in someone’s head, it isn’t usable data yet. Put it in a transcript, a note, or a database, and the system has something to work from.
A client file on your desk is the everyday version. Some of it is neat, like the signed agreement and the invoice. Some is messy, like handwritten notes and a transcribed voicemail. It’s all information, but some pieces are much easier to use than others.
Why you care
You can start with raw material: a model can read a messy transcript, pull out themes, and turn it into structured data. Messy still costs you, though. If the source is incomplete, stale, or scattered across six places, the model will struggle. The practical move is to make the work visible: record the call, save the note, name the source clearly. AI can’t help with work it can’t see.