What the manager sees
Planning in grams of metal
Weight is entered by week, and the currency value is derived from the price per gram. The natural unit comes first, money comes second.
Home · Case studies · Jewelry manufacturing
Case study · Jewelry manufacturing
Forecasting, metal pricing and production — in one system
For a full-cycle jewelry manufacturer, we built a platform that plans sales in grams of metal rather than in currency, revalues items daily against exchange quotes, and links the commercial plan to the production plan.
52under-shipped SKUs analyzed by AI
6/6weeks of forecast filled in
19customers under delivery control
82,5%order fulfillment for a specific SKU
001
Why a standard planning approach didn’t fit
A full-cycle jewelry manufacturer: proprietary items, precious metal and gemstones in the cost base, several sales channels. Standard planning tools based on currency and unit counts don’t fit this specificity — so we built an industry-specific platform that unifies forecasting, sales, pricing and production in one loop.
01
The planning unit is not currency
The base unit of the business is grams of precious metal. Money is a derived value, not the primary one.
02
The metal price changes every day
Cost and the value of finished items depend on current exchange quotes, not on the price at the moment of data entry.
03
The commercial and production plans were separate
The sales plan and the production plan were maintained separately and were not linked.
Forecasting
→
Sales
→
Pricing
→
Production planning
Specifics
What the system accounts for in jewelry manufacturing
What happens daily
Accounting for gold and gemstones, daily revaluation
Cost is built up from the value of metal and gemstones, and the value of finished items is recalculated every day against exchange quotes.
What’s visible in the forecast
Sales channels linked to production
Forecasting is broken down by sales channel, and the sales plan and the production plan are visible on a single screen and calculated from the same unit of measurement.
Results
What the system shows today
An example from the AI assistant’s work: a breakdown of under-shipped items for one month, and how completely the August forecast was filled in.
Found by the assistant
52
Under-shipped items for July, broken down by series and SKU.
Potential lost revenue
≈RUB 5.3M
An estimate for these 52 items, calculated in code from orders and shipments in 1C.
Forecast completeness
6 / 6
Weeks of August covered by the forecast — no empty “customer × category” pairs.
Customers affected
19
Customers whose orders were included in the month’s under-shipment analysis.
In the interface
Forecasting and control on one screen
The manager enters weight by week — the system calculates the currency value from the price per gram automatically. Management sees plan coverage and forecast completeness before the quarter ends.
udev.forecast · Forecasts · MY FORECAST vs MANAGEMENT PLAN
udev.forecast · Dashboard · Q3 2026
Input in a natural unit
Weight by week
Plan coverage
66% of the management plan
Completeness
6 of 6 weeks in August
Plans are shown (my forecast and the management plan), not actual sales. The weekly grid is cut along month boundaries, not by the calendar week.
AI assistant
Lost revenue traced to a specific order
A question in plain words turns into a breakdown: how much was ordered, how much was shipped, where the losses are and why. Any figure can be drilled down to the item code, customer and order number.
01
Analysis of lost sales
→
02
Calculation of potential lost profit
→
03
Identifying opportunities
→
04
Supporting commercial decisions
udev.forecast · AI assistant · Data from 1C
Series 03-2902 and 03-2901 — the main risk area75 unshipped items across the two series: 28.3% of all under-shipments and 25% of total lost revenue.
9 of 52 SKUs with zero shipments5 of them have already been discontinued but are still in demand.
The principle of trusting the numbers
The figures are calculated in code from orders and shipments in the accounting system — the assistant only explains the numbers. Every conclusion shows the underlying document numbers, down to a specific SKU and order.
One system instead of scattered spreadsheets
The company got a connected management system: data, forecasting, sales, pricing and production planning — on one platform and in the units of the real business.