--- title: Know how much each product will sell next month. description: Your AI reads the ERP's sales history, works out each product's moving average and seasonality and saves next month's forecast, with every number's source. language: en-US canonical: https://www.siglata.com/en-US/trabalhos/previsao-de-demanda --- # Know how much each product will sell next month. Canonical: https://www.siglata.com/en-US/trabalhos/previsao-de-demanda Your AI reads the sales history exported from the ERP and saves each product's forecast, with the math and the source of every number. ## The month's purchase comes from a guess The sales history is in the ERP, but the forecast comes from a spreadsheet with numbers copied from last month. Nobody looks at what happened in the same month last year, and the product that sells more in October runs out again. ## One question, the forecast per product You ask how much will sell in October. Your AI adds up each product's last three months, applies last October's index and saves the forecast as a new version of the spreadsheet, with the previous one kept. ## Every number with its source Tap a forecast to see the three-month average, the October index and the row of every number in the history. ## The forecast method is written down How many months go into the average and which index applies to each month live in a company skill. Each person works with their own access, and every change stays in the history, with the previous version kept. 1. The month's purchase comes from a guess: The sales history is in the ERP, but the forecast comes from a spreadsheet with numbers copied from last month. Nobody looks at what happened in the same month last year, and the product that sells more in October runs out again. 2. Know how much each product will sell next month.: Your AI reads the sales history exported from the ERP and saves each product's forecast, with the math and the source of every number. 3. One question, the forecast per product: You ask how much will sell in October. Your AI adds up each product's last three months, applies last October's index and saves the forecast as a new version of the spreadsheet, with the previous one kept. 4. Every number with its source: Tap a forecast to see the three-month average, the October index and the row of every number in the history. | Product | Jul | Aug | Sep | 3-month average | October index | October forecast | Sales/day | | --- | --- | --- | --- | --- | --- | --- | --- | | Café 500 g | 150 | 155 | 160 | 155 | 1.20 | 186 | 6 | | Azeite 500 ml | 120 | 124 | 128 | 124 | 1.00 | 124 | 4 | | Leite 1 L | 640 | 620 | 600 | 620 | 0.90 | 558 | 18 | | Açúcar 1 kg | 320 | 310 | 300 | 310 | 0.90 | 279 | 9 | | Detergente 500 ml | 126 | 124 | 122 | 124 | 1.00 | 124 | 4 | | Óleo 900 ml | 190 | 186 | 182 | 186 | 1.00 | 186 | 6 | | Ovos 12 un | 285 | 279 | 273 | 279 | 1.00 | 279 | 9 | | Manteiga 200 g | 240 | 248 | 256 | 248 | 1.00 | 248 | 8 | | Feijão 1 kg | 210 | 217 | 224 | 217 | 1.00 | 217 | 7 | | Molho 340 g | 90 | 93 | 96 | 93 | 1.00 | 93 | 3 | | Arroz 5 kg | 300 | 310 | 320 | 310 | 1.00 | 310 | 10 | | Biscoito 400 g | 118 | 124 | 130 | 124 | 1.25 | 155 | 5 | | Total | — | — | — | — | — | 2,759 | 89 | ### Biscoito 400 g · October forecast 155 2026 sales: July 118, August 124, September 130 units. 3-month moving average: (118 + 124 + 130) ÷ 3 = 372 ÷ 3 = 124. October index: October 2025 sold 120; July to September 2025 sold 90, 96 and 102, an average of 96. 120 ÷ 96 = 1.25. October forecast: 124 × 1.25 = 155 units. Sales per day: 155 ÷ 31 days = 5. Vendas_ERP_jul2025_set2026.csv v1: Biscoito 400 g rows from Jul 1 to Sep 30, 2026 · sums: 118, 124 and 130 Vendas_ERP_jul2025_set2026.csv v1: Biscoito 400 g rows from Jul 1 to Oct 31, 2025 · sums: 90, 96, 102 and 120 Previsao_de_demanda : (company skill) · item 1: forecast = average of the last 3 months × the same month's index from the year before | | | | In October 2025, the cookies sold 25% above the average of the three months before. The 2026 average alone would give 124; with the index, the forecast is 155. Previsao_de_demanda.xlsx v3 to v4 Product · 3-month average · Jul to Sep 2025 (average) · Oct 2025 · Index · Forecast · Sales/day | Product | 3-month average | Jul to Sep 2025 (average) | Oct 2025 | Index | Forecast | Sales/day | | --- | --- | --- | --- | --- | --- | --- | | Café 500 g | 465 ÷ 3 = 155 | 150 | 180 | 180 ÷ 150 = 1.20 | 155 × 1.20 = 186 | 186 ÷ 31 = 6 | | Azeite 500 ml | 372 ÷ 3 = 124 | 110 | 110 | 1.00 | 124 | 124 ÷ 31 = 4 | | Leite 1 L | 1,860 ÷ 3 = 620 | 600 | 540 | 540 ÷ 600 = 0.90 | 620 × 0.90 = 558 | 558 ÷ 31 = 18 | | Açúcar 1 kg | 930 ÷ 3 = 310 | 290 | 261 | 261 ÷ 290 = 0.90 | 310 × 0.90 = 279 | 279 ÷ 31 = 9 | | Detergente 500 ml | 372 ÷ 3 = 124 | 115 | 115 | 1.00 | 124 | 124 ÷ 31 = 4 | | Óleo 900 ml | 558 ÷ 3 = 186 | 170 | 170 | 1.00 | 186 | 186 ÷ 31 = 6 | | Ovos 12 un | 837 ÷ 3 = 279 | 260 | 260 | 1.00 | 279 | 279 ÷ 31 = 9 | | Manteiga 200 g | 744 ÷ 3 = 248 | 230 | 230 | 1.00 | 248 | 248 ÷ 31 = 8 | | Feijão 1 kg | 651 ÷ 3 = 217 | 200 | 200 | 1.00 | 217 | 217 ÷ 31 = 7 | | Molho 340 g | 279 ÷ 3 = 93 | 80 | 80 | 1.00 | 93 | 93 ÷ 31 = 3 | | Arroz 5 kg | 930 ÷ 3 = 310 | 290 | 290 | 1.00 | 310 | 310 ÷ 31 = 10 | | Biscoito 400 g | 372 ÷ 3 = 124 | 96 | 120 | 120 ÷ 96 = 1.25 | 124 × 1.25 = 155 | 155 ÷ 31 = 5 | [Download the example (.xlsx)](https://www.siglata.com/en-US/trabalhos/previsao-de-demanda/planilha.xlsx) Illustrative data. [With the forecast saved, your AI works out what to restock.](https://www.siglata.com/en-US/trabalhos/reposicao-de-estoque) ## Frequently asked questions ### Which ChatGPT does it work in? ChatGPT Business and Enterprise: an admin enables Siglata once for the team. On Pro, Siglata is read-only, in developer mode. On Plus and Free, it doesn't work in ChatGPT on the web. In Codex, each person connects it, on every plan that includes Codex. ### Does it work with my ERP? Two ways: with the ERP's exports, as spreadsheets or CSV, and in the ERP's own pages, like Bling: your AI signs in to the portal with the login you or the company connected, reads lists and tables, clicks, fills in and posts, without seeing the password. The sales history by product works the same way, as a spreadsheet or CSV. ### How do I calculate a demand forecast? The simplest way is the moving average: add up the last three months' sales and divide by three. If the month is seasonal, multiply by the same month's index from the year before, which is that month's sales divided by the average of the three months before it. With an average of 124 and an index of 1.25, the forecast is 155. ### How do I calculate a forecast in Excel? In Excel, the three-month moving average is =AVERAGE(B2:D2), and the forecast with the index is =AVERAGE(B2:D2)*F2. Excel also has Forecast Sheet, on the Data tab (Microsoft Support, read on Oct 6, 2026). Your AI does the same math straight on the ERP export and saves the spreadsheet with the source of every number. ## Want to see it with your company's history? We'll show it with your ERP's sales export, in a WhatsApp conversation. [Let's talk](https://wa.me/5516991184410?text=Hi!%20I'd%20like%20to%20see%20demand%20forecasting%20with%20Siglata.)