AI Agents for a Fine Jewelry Brand

Key results

$81k-$162k per year

Additional custom-order revenue, before costs

~2,000 hours per year

Hours of routine work the agents take off four roles, before time spent reviewing their output

78% → 95%

Share of custom inquiries that receive a quote

15 → 5

Minutes per order

jewelry scheme

Summary

The brand is a US fine jewelry business that sells on Shopify Plus and has over $10 million in annual revenue. Each of its custom rings needs a consultant, a stone supplier and a workshop, while the information sits in separate places: customer messages in Gorgias, shipping in ShipStation and production updates in a shared spreadsheet. There is no in-house developer to tie them together.
We connected that work through four AI agents that collect missing details, prepare quotes and order reviews, draft price changes and track service requests. The team stays in its existing tools and approves anything that commits money, inventory or a delivery date.
Cooperation Period
Ongoing
Location
USA
Industry
Jewelry
Service Provided
AI Agents
jewelry showroom

Business Challenge

A customer can be ready to buy while the team is still chasing the details.

  • Custom inquiries go cold. Consultants collect specs across several messages, then check stones with suppliers before quoting.
  • High-value orders wait. The COO gathers payment, customer and shipping data manually before deciding whether to ship.
  • Prices lag the metal market. Repricing is a monthly spreadsheet pass, so when gold moves, the store keeps selling at old prices for weeks. Lab-grown and treatment disclosures also vary across listings.
  • Service requests lose their trail. Resizing and repair requests move between the inbox and the workshop with no clear owner.

What We Did

Four agents:

  • 1

    Custom Order Desk

    Picks up custom inquiries from the site, email, chat and Instagram, asks for the missing specs and drafts quote options with matching stones for a consultant. It checks each stone’s availability with the supplier before quoting and again at deposit. After the customer approves the CAD, the desk tracks casting, setting and QC as the workshop confirms each stage.

  • 2

    High-Value Order Review

    When Shopify Flow flags an order over $2,500, it builds one review packet from customer history, payment risk signals and shipping details. The COO decides whether to ship, hold or cancel, and approved orders ship insured with a signature required. If a chargeback comes in, the agent assembles the evidence for the dispute.

  • 3

    Catalog and Pricing

    Drafts listings from grading-report data and flags lab-grown and treatment disclosure gaps. It checks the gold price daily and, after a 3% move, prepares a repricing batch for every variant with the margin impact. A merchandiser approves every listing and price change; if report data is missing, the field stays empty and the listing waits.

  • 4

    Post-Purchase Service

    Answers order-status questions in Gorgias from live Shopify and carrier data. Resize, repair and return requests become tracked cases, and staff approve costs and refunds after inspection. Appraisal requests go to a qualified partner, and the agent never states value or authenticity. Everything else gets a drafted reply for a rep to send.

Results

More time for custom-ring customers
Less preparation time per inquiry. Consultants review a complete brief and quote options instead of rebuilding the request across inboxes.
High-value orders reach a decision faster
A short review of a prepared packet replaces a manual check. Insurance, signature requirements and any exceptions are confirmed before release.
Prices keep pace with metal costs
Repricing no longer waits for a monthly pass. Merchandisers see which variants need attention and how each change affects margin.
Customers get a traceable next step
Service and production requests carry an owner, a confirmed status and the next action. Staff can answer without reconstructing the order history.