← ALL CASE STUDIES
Case Study · Retail & Jewelry

Nine years of manual research. Done in hours.

Every growing retailer has one: the backlog everyone stopped mentioning. For this jewelry business it was nine years of acquisitions — thousands of items — waiting for research nobody would ever have time to do. Then it took an afternoon.

9y→hrsProcessing time, years to hours
11,400+Items identified, graded, priced
100%Cataloged with provenance notes
Jewelry retailer & estate buyer
IndustryRetail & Jewelry
InventoryEstate acquisitions · vintage · loose stones
Built byPromata — Blueprint → AI Build
StatusComplete · engine now runs on every new acquisition

Does this sound like your business?

There’s a backlog in your operation that everyone has agreed to stop mentioning.
“We’ll get to it in the slow season” — said for the ninth consecutive year.
Items sell below their value because nobody had time to research what they actually are.
Exactly one person knows how to price the difficult pieces.
Real capital is sitting frozen in drawers, waiting for paperwork.
IF YOU CHECKED THREE OF THESE — THIS CASE STUDY IS ABOUT YOU.
The Challenge

The backlog that ate nine years.

Estate lots arrive as boxes: pieces, loose stones, old invoices, sometimes an appraisal from 1987. Turning a box into sellable inventory meant expert research — identify the piece, verify materials, grade condition, establish provenance, price against the market. At twenty-plus minutes per item, the arithmetic was brutal: the backlog grew faster than any human could shrink it.

⚠ THE BACKLOG ROOM
11,400 items across nine years of boxesUNRESEARCHED
~20 minutes of expert research per item3,800 HOURS NEEDED
“Winter project” — attempted, abandoned9 YEARS RUNNING
Unknown pieces sold at estimateVALUE LEFT ON TABLE
One expert who can price the hard piecesSINGLE POINT OF FAILURE
✦ THE RESEARCH ENGINE
Photo + invoice + appraisal read together, per itemSECONDS EACH
Identified, graded, priced against live market dataAUTOMATICALLY
The whole archive processedONE AFTERNOON
Every item listed with provenance notesFULL VALUE
Expert reviews only the flagged piecesJUDGMENT, NOT VOLUME
VS

THE SAME 11,400 ITEMS — 3,800 HOURS OF RESEARCH VS. ONE AFTERNOON

  • Research was the bottleneck, not selling

    Demand existed. Shelf space existed. What didn't exist was 3,800 expert-hours — so nine years of acquisitions sat as cost instead of inventory.

  • Unknown items leak value silently

    A piece sold without research sells at a guess. Across thousands of items, the guesses compound into a number nobody wanted to calculate.

  • Expertise trapped in one head

    The owner's eye for the difficult pieces was the business's greatest asset and its greatest constraint — it didn't scale, and it couldn't take a vacation.

  • Every new lot deepened the hole

    Acquisitions kept arriving. The backlog wasn't a project; it was a ratchet.

The Solution

An engine with the expert’s eye — at machine speed.

We built a research pipeline that reads everything the business already had — item photographs, purchase invoices, old appraisals — and does what the expert does: identify, verify, grade, price, and write it up. The expert’s knowledge trained the rules; the engine applies them to thousands of items without fatigue.

Estate lot #47 — 212 items, photos + 1987 appraisalScanned this morning
IDENTIFIES · GRADES · PRICES
Visual identificationMaker's marks · cuts · settings
Market & auction dataLive comparable pricing
Inventory systemListed with provenance notes
212 ITEMS CATALOGED BEFORE LUNCH · UNCERTAIN PIECES FLAGGED FOR THE EXPERT'S EYE
  • Reads the whole paper trail

    Photos, handwritten invoices, decades-old appraisals — cross-referenced per item, so the catalog entry carries the story, not just a description. Provenance is value in this trade.

  • Confidence-scored, expert-reviewed

    The engine knows when it's guessing. High-confidence items flow straight to listing; ambiguous marks and unusual pieces queue for the owner — whose time now goes exactly where it's irreplaceable.

  • Priced against the living market

    Comparables from current market and auction data, refreshed continuously — not the sticker instinct of a busy afternoon.

  • Permanent, not one-off

    The same pipeline now runs on every new acquisition. The backlog isn't just cleared; the concept of a backlog is gone.

Results & Business Impact

Nine years of frozen capital, thawed in an afternoon.

  • The entire archive became sellable inventory

    Nine years of acquisitions moved from boxes to listings — researched, priced, and documented to a standard the manual process never had time for.

  • Value recovered on the “unknown” pieces

    Items that would have sold at estimate now sell with identification and provenance — the difference, across thousands of items, funded the build many times over.

  • The expert became a reviewer

    From twenty minutes per item to seconds per flagged item. The rarest resource in the business now touches only the pieces that deserve it.

  • New lots are cataloged on arrival

    The nine-year problem can't happen again — research now happens at the speed of acquisition.

The backlog was never a discipline problem.

No one was lazy for nine years — the math was simply impossible for humans. That’s the mark of work that belongs to machines: infinitely repetitive, rule-describable, endlessly postponed. If your operation has a drawer, a folder, or a warehouse everyone has agreed not to discuss, that’s not shame. That’s your Blueprint.

9y→hrsProcessing time
11,400+Items cataloged
0Items skipped

What we didn't solve

Attribution is only as good as the piece allows. Items with no maker's mark, illegible hallmarks or ambiguous provenance are catalogued with that ambiguity recorded rather than resolved, and a specialist reviews them before they are priced.

What's sitting in YOUR backlog?

Every business has one. Tell us about yours — we'll tell you what clearing it is worth, before you spend a dollar.

Book a discovery call

Want results like these?

Tell us where your team is losing hours. We'll show you exactly what automation can do about it — and what it's worth, before you spend a dollar.

30 minutes · No pitch deck · An honest first read

Case study: nine years of manual research, done in hours