Who this is for
- Distributors and retail chains deciding reorders by feel, with cash stuck in slow stock and gaps in fast stock.
- Manufacturers planning production and raw-material purchases from last month's number and a guess.
- Owners with receivables ageing who want to know which accounts to chase this week, not after they go bad.
- Firms that already have a dashboard and now want it to say what to do next, not just what happened.
- Businesses with a Kerala and a UAE operation that want one forecast across both, in rupees and dirhams.
What we build
- Demand forecasts by item, branch and week or month, built from your sales history, seasonality and known events such as Onam or Ramadan.
- Reorder suggestions that account for lead time, minimum order quantity, current stock and the forecast, shown as a list your purchase team can accept or edit.
- Receivables risk flags: customers whose payment behaviour has changed, ranked for follow-up with the reason shown.
- Scenario models: what if we raise prices on this range, open a branch, or drop a slow line; the model shows the expected effect on sales, margin and cash.
- Margin leak detection: discounts, freight, returns or rework that are eating margin on specific customers or products.
- Anomaly alerts: a sudden drop in a rep's sales, a supplier price jump, or a stock count that does not match the system.
- Delivery in the places you already look: your dashboard, a morning WhatsApp or email summary, or a sheet for the purchase team.
Typical scope
A pilot usually covers one decision for one part of the business, most often reorder suggestions for the top-selling items in one branch or warehouse, or payment-risk flags for the receivables book. We start by testing the model on your past data and showing you the error honestly. If it is not better than what you do today, we stop there and you pay for the pilot work done, not the full build. This is not a black box and not a tool for messy data: every suggestion carries its reason, and the model is only as good as the entries in Tally or your ERP. If your sales history is short, lives on paper, or has items under several names, decision support is the wrong first step; a dashboard or Discovery Sprint will give you more for less, and we will say so.
What you receive
- Working models running on your own data, hosted in your cloud account, refreshed on a schedule.
- A back-test report showing how the forecast would have done against your real past months, so you know its error before you rely on it.
- The suggestions and flags delivered into your dashboard, WhatsApp summary or sheets.
- A plain-English note on how each model works, what it uses and what it ignores.
- A data-quality fix list for anything that limited accuracy, with who should fix it.
- A measurement sheet comparing the agreed success metric before and after the pilot.