Documentation Index

Fetch the complete documentation index at: https://docs.wair.ai/llms.txt

Use this file to discover all available pages before exploring further.

Highlevel interface architecture

Prev Next

We advise to look into the Integration core principles after you have read this page and corresponding documentation.

Below is a high level flowchart of the interface architecture. We have a public API where you need to push your data to, and pull our output (recommendations) from, to process back in your source system. The source system is typically an ERP system.

We have different products with different features, below is an overview of the data required as input and what is generated as output from Wair:

Product

Feature

Input needed

Wair Output

Note

Suzie

Product descriptions

Item + item masterdata + images

Product description and product description tags

Wallie

Replenishment

Item + item masterdata + images + stock + sales + transfers (+ stock targets) + locations

Either stock targets or transfers

If an ERP system has stock target (aka Min/Max aka Reorder point) based replenishment logic that already suits the customers' needs, we generate intelligent stock targets for the ERP and rely on the ERP to create the transfers. If this is not the case we can do the replenishment engine in Wair en generate transfers

Wallie

Redistribution

Item + item masterdata + images + stock + sales + transfers + locations

Transfers

Wallie

Initial distribution

Item + item masterdata + images + stock + sales + transfers + purchases + locations

Transfers

Below is a scope per data set which is required. There is also a link to the API documentation, which always contains best practices, use cases, common pitfalls and a more detailed explanation how the data is used for reference:

Data set

Scope

Documentation

Items and item masterdata

We recommend to send the product information of all the tradeable goods. We do not recommend to filter out outdated products, because we ask for historical sales/stock, which we need to match with product information. See documentation for more details.

POST /Items

Images

We also recommend to send all your images, so we can match them to all the products. See documentation for more details.

POST /Images /

POST /Images/urls  

Stock

Both current and historical stock data are required. Historical stock must be provided as mutations. Without historical stock, especially for locations that sell goods, backtesting is not possible. See the documentation for more details.

POST /StockMutations  /

POST /Stock

Sales

Historical sales data is required to train the AI model. Only sales to end consumers need to be shared. We recommend at least two full calendar years; more data is always better and can improve model accuracy. Fewer than two years is still usable but may impact prediction accuracy. See the documentation for more details.

POST /Sales

Transfers

By transfers, we mean all upcoming stock movements. This includes store-to-store, warehouse-to-warehouse, and warehouse-to-store transfers, as well as non-Wair–managed orders that move goods from a warehouse (for example, B2B sales orders). Historical transfer data is not required. See the documentation for more details.

POST /Transfers

Purchases

We need to have the open purchase orders, no historical purchase order data is needed. See documentation for more details.

POST /PurchaseOrders

Locations

All locations that have (had) stock and/or are selling to end consumers. See documentation for more details.

POST /Locations