AI Document Utility: Scan, Extract, Rename, File — One Pipeline
Document AI is a workflow, not a chatbot. Scan → fields → a sane filename is the job scanners skip.
Object-ID apps are rising. The leftover job is a cautious ID plus a range — not a fake appraisal certificate.
Antique and object-identifier apps are taking niche reviews because flea-market and inheritance photos have nowhere trusted to go. Lens shops SKUs. Users want a name and era, then a humble price band.
This brief is part of the Android problem-space map. It is here so we can measure interest — opens, read depth, and related clicks — before anyone writes an app spec.
Google Lens, Facebook “what is this,” Reddit, a dealer WhatsApp. Wrong century IDs spread. People screenshot a $12,000 eBay ask and believe it.
Lens, eBay, specialist dealer apps, viral “identify anything” AI. Trust is scarce. Hallucinated maker marks will get you one-starred and, in some regions, legal trouble.
Photo → candidate names + period + similar sold comps + “we are not an appraisal.” One category first (ceramics, watches, or tools). Human-in-the-loop later. Never mint a certificate. Offline field notes for markets with bad data.
Show ten inheritance photos to a dealer and to your prototype. Score name accuracy, not wow. If Lens already names it, walk away. If users need “don’t get scammed at the market,” you have a job — and a duty to be boring.
Read the full research map to compare solo potential across all forty spaces.