AguilarTech — Procurement AI Agent Demo

💡 About this demo — what you're looking at & how to interactAGENTIC AI SHOWCASE

▼

Simulate a working day

Press ▶ Next Day to advance the simulation. Each day, inventory is consumed by demand, in-transit orders move closer to arrival, and the AI agent wakes up to analyse the entire network.

Feed the agent unstructured intel

RAG

The inbox simulates real operational emails. Select one before pressing Next Day and the agent applies Retrieval-Augmented Generation (RAG) — extracting structured supply chain signals (demand spikes, delay durations, affected SKUs and nodes) directly from the plain-text email body, then injecting those values into the inventory mathematics.

Watch agentic reasoning live

ReAct Pattern

The Agent Reasoning Log demonstrates the ReAct pattern (Reasoning + Acting) in real time: the agent reasons about the network state, calls a tool to simulate Days of Stock, observes the result, reasons about what action is needed, calls another tool — iterating until every high-velocity SKU has been checked and all stockout risks have been addressed.

You stay in control — always

Human-in-the-Loop

Every recommendation lands in Agent Recommended Actions. This is Human-in-the-Loop (HITL) governance — a core principle of responsible agentic AI design. Nothing is committed to the network until you explicitly approve it. You can approve, reject, or simply ignore any proposed action.

📧 How the inbox works: Each email simulates a real operational message a demand planner might receive. Click any email to expand it, then select the one you want the agent to analyse. When you press ▶ Next Day, the agent applies Retrieval-Augmented Generation (RAG) to extract the relevant supply chain signals from the plain text — demand multipliers, delay durations, affected products and locations. Each email can only be processed once. Use the ✏ edit and 🗑 delete buttons to customise messages, or add your own to test the agent with any scenario you like.

⚠️

SKU locations — low stock

🚫

Locations stocked out

🚚

Active orders in transit

⏳

Actions awaiting approval

Days of Stock — All Locations & Products

Days of stock remaining and on-hand unit quantities at each warehouse, factoring in current demand rates and in-transit orders

>7d — OK

3–7d — Low

<3d — Critical

Out

Product	Type	MEL	SYD	PER	AKL

📧 Inbox

Select an email before pressing Next Day — the agent will extract intel from it using RAG

How it works: Click an email to expand it. Select it before Next Day to trigger Retrieval-Augmented Generation (RAG) — the agent reads the plain text and extracts supply chain signals automatically. Use ✏ / 🗑 to edit or delete. Add your own test scenarios below.

From

Subject

Message body

Agent Reasoning Log

Real-time view of the ReAct loop — tool calls, observations & decisions

Agent reasoning will stream here.
Select an email → then press ▶ Next Day