CASE STUDY / 001

AI Procurement
Workflow.

Automating the transition from an unstructured procurement request to a structured inventory decision.

PythonWorkflow automationCSVAPI example

PROJECT STATUSPublic repositoryExplore source code ↗

The problem

Procurement requests often arrive as free-form language. Staff must interpret the request, check stock and decide what action to take. Repeating these checks manually is slow and prone to inconsistencies.

How it works

Receive requestAccept a natural-language procurement message.
Extract informationIdentify requested item, quantity and location.
Check inventoryCompare the request against structured inventory records.
Recommend an actionReturn a structured result, such as fulfil from stock or find a supplier.

Example API interface

POST /run-workflow
{
  "message": "Need 500kg fertiliser for Canterbury farm next Wednesday"
}

// Example response from the README
{
  "item": "fertiliser",
  "quantity": 500,
  "location": "Canterbury",
  "inventory_status": "available",
  "recommended_action": "fulfill_from_inventory"
}

The example is taken from the project README and is illustrative rather than a live API response.

What I would improve next

Introduce automated tests, more robust extraction validation, structured error handling and execution metrics. These are improvement opportunities, not features claimed to be complete.

View repository ↗All projects ↗