Inventory confirmed by asking
Real availability depends on calling the warehouse because movements, reservations and returns do not reconcile.
We connect inventory, purchasing, sales, invoicing and costs on clean master data, clear rules and workflows the team can run every day.
| Code | Item | Stock | Status |
|---|---|---|---|
| ART-1042 | Folding box 40 × 30 | 318 pcs min. 80 | OK |
| ART-1188 | Stretch film 500 mm | 12 pcs min. 60 | Low stock |
| ART-1205 | Thermal label 10 × 15 | 2,400 pcs min. 500 | OK |
| ART-1311 | Packing tape 48 mm | 146 pcs min. 40 | OK |
Purchasing, Sales and Finance read that same record: one source of truth.
We build on what your team already knows. If the platform has an API or webhooks, it integrates; if it does not, we look for the most stable path before proposing a change of tool.
And any other platform with a public API.
Stock lived in three separate spreadsheets, purchases were written down in a notebook, and the invoicing system talked to neither. We brought the ERP up module by module, in an order decided beforehand, and only switched the old one off once the new one closed a full month with no discrepancies.
Furniture manufacturer · Argentina
The data came from three stock spreadsheets, an invoicing system more than ten years old and the purchasing notebook. The first job was unifying the item master: the same fitting was on file under four different codes depending on who had written it down. Balances were checked against a physical count before they went in. Only the last two financial years of history were migrated: anything older stayed readable in the old system, rather than dragging in years of data nobody was going to look at.
Stock went live, and purchasing went live on top of stock. Production only came in once those two were steady, and costing at the close of the quarter. Turning on every module on the same Monday is the classic mistake: if something breaks, nobody knows which of the six broke it, and the team loses faith in the whole system over one screen that fails.
Invoicing draws down stock on the spot, with nobody keying the movement separately. When an item hits its reorder point, the system raises the purchase request with the usual supplier already suggested. And when a purchase invoice arrives at a different price, the item cost updates on its own and carries through to the finished product cost. There is one stock figure for the warehouse, sales and production, instead of one per area.
Training per role and with the plant’s own data, not with a textbook example: the warehouse lead learns to receive goods, not to use an ERP. Then a month in parallel, keying into both systems and comparing the daily closes side by side. For the first few weeks three things were checked every morning: gaps between system stock and warehouse stock, invoices rejected by the tax authority, and purchase requests the ERP raised that the lead could not explain.
The same mechanics adapt to a parts importer, a food plant or a distributor: the modules that go first and the reorder rules change, the way of starting does not.
When purchasing, sales, warehouse and finance work from different files, every figure needs an explanation and no decision starts from the same base.
Real availability depends on calling the warehouse because movements, reservations and returns do not reconcile.
Prices, expenses and product composition are rebuilt at month end, after the sale has already happened.
The same product, supplier or customer appears under different codes and breaks reporting and integrations.
The goal is not to activate every module. It is to build a reliable sequence from source data to the journal, inventory or report that needs it.
Products, price lists, customers, suppliers, warehouses and taxes with defined codes and owners.
Request, approval, order, receipt and variance connected to inventory and accounts.
Order, reservation, delivery, invoice and margin using the same data chain.
Connections to ecommerce, CRM, banks or other platforms, with alerts when a workflow remains incomplete.
A total, simultaneous cutover concentrates too much risk. We prioritise the highest-impact flow and use that delivery to validate master data, roles and integrations.
We follow documents and data across teams to find where they are duplicated, corrected or lost.
We define structure, ownership, creation rules and cleaning before loading the system.
Test environment, real scenarios, validated balances and permissions for each role.
Operational support, variance control and expansion once the first chain reconciles.
Configuration and migration are only part of the job. Rules, ownership and controls keep the data reliable after go-live.
Yes. We choose a platform based on processes, localisation, integrations, internal capacity and total cost of ownership—not feature count alone.
Not necessarily. Sometimes the right answer is to retain the transaction system and integrate the missing piece. Discovery separates configuration, data and infrastructure problems.
Yes, with a cutover date, transformation rules and validation against known totals. History is migrated only when it has a concrete use.
It depends on workflows, master-data quality, localisation and integrations. We propose phases with usable releases rather than one long project with no intermediate result.
Only when the difference represents a real advantage or obligation. If the standard process works without losing value, we avoid code that must be maintained later.
We trace the data to its source and determine whether the problem needs configuration, integration or replacement of one part of the system.
Kai Kommit agent · Online
Hi — I'm Kai, Kommit's agent. Tell me which process is costing you hours and I'll tell you where to start.
AI agent: it can be wrong. Anything agreed is confirmed by the team. Prefer the form →