Rami Daud
Case file 01 · AdoptKent State UCM · fall 2026 · status: in progress

Case file 01 · Adopt

The Lightroom archive workflow

Claude suggests. A photographer decides. Nothing reaches PhotoShelter without a human review.

Who decides
A photographer approves, edits or rejects every suggestion.
Compared against
PhotoShelter’s own AI features and what they cost.
Measured next
Time saved, once it’s part of regular filing. No savings claimed until then.

The problem

A university photo archive is only as useful as its metadata. When someone across campus searches for a picture, what they find depends on whether the keywords and descriptions were written well, sometimes years ago. Writing that metadata by hand is the slow part.

What I did

This fall I connected Claude to Lightroom Classic to check and strengthen keywords, descriptions and the team's five-star rating scale. I started with the last two months of work and I'm moving backward through the catalog.

Flight plan

  1. Shoot lands in Lightroom Classic

  2. Claude suggests keywords and descriptions

  3. A photographer approves, edits or rejects

  4. Published to PhotoShelter and Smart Galleries

Airtable stays the source of truth for every assignment.

Results so far

Thousands of images worked through, and still moving back through the catalog. Time saved will be measured once this is part of how we file new shoots, not a cleanup project.

What it means for a team

Find a real bottleneck first, keep a person accountable for the result, and have numbers in hand before asking anyone else to adopt it. I'm also weighing it against PhotoShelter's own AI features and their cost, because sometimes the right answer is a tool you already pay for.

How a suggestion is judged

  • Accurate. It describes only what is in the frame.
  • Verified. Names, places and events match the assignment record in Airtable.
  • No guessing at people. Nobody is identified unless the photographer already identified them.
  • Findable. Keywords match the words people across campus actually search for.

A suggestion that misses any one of these is edited or rejected.

What I will measure

  • How many suggestions are approved as written, edited, or rejected
  • Minutes from a shoot landing to being searchable in PhotoShelter
  • Whether colleagues find what they search for, before and after

Until those numbers exist, this page claims no time savings.

Who is involved

I run the workflow as the team's digital asset lead and review every change before it publishes. Before asking anyone else on the team to use it, I'll have the numbers above.