AI Demand Forecast

Weekly demand forecasting: what Chef7's AI actually looks at

Chef7's demand forecast crosses your last 30 days of served recipes with current stock to project next week's consumption. It doesn't use weather data or external events — it reads your own operating history. The goal: anticipate supplier orders and flag ingredients about to run out.

What the AI reads

• 30-day recipe history: main ingredient, category and served date
• Current stock in warehouse and kitchen (normalised to kg, L or units)
• Usage frequency per ingredient over the last 30 days
• Stock coverage: how many days the current balance lasts at average consumption

What the dashboard shows

1. Top ingredients used: bar chart with the 10 most-consumed over the period
2. Weekly forecast table: ingredient, predicted use (per week), current stock, suggested action (OK, Buy, Urgent)
3. Stock at risk: items whose current stock falls below weekly predicted use
4. One-click shopping list: generates a supplier-ready list with quantities pre-calculated

Illustrative example

Imagine a traditional restaurant with a fish-and-meat menu. On a Friday, the dashboard might show 48 recipes served in the last 30 days, with cod, octopus, onion and potato at the top by volume. If cod and seabass have stock below weekly expected use, they're tagged "Urgent"; a single click on "Generate Shopping List" prepares the supplier order with quantities pre-filled.

When the forecast matures

In the first 7-14 days, history is thin and predictions are conservative. Forecast quality stabilises after 30 days of operation. Every served recipe feeds the model — the AI learns from your own data, no external feeds required. For context: the National Restaurant Association (State of the Restaurant Industry 2024) reports back-office automation typically reduces cost of goods sold by 3 to 5%.

Source: National Restaurant Association (2024).