How I Connected AI Agents with Manufacturing Supply Chain

Source: Dev.to

My Project in 34 Words

An enterprise AI‑agent system that monitors component prices across suppliers for IoT manufacturing companies. The system helps maintain profit margins on tender commitments by detecting price spikes before they impact production costs.

Background Issue — The Manufacturing Business Challenge

Manufacturing projects span months and involve multiple phases: tender submission, design, testing, evaluation, and mass production. Companies must commit to prices during the tender phase, often months before actual production begins.

The Risk

A company may win a tender for smart‑TV controllers at $5,000 each. If, months later, the price of a critical component (e.g., the microcontroller) spikes twice, the original budget is breached, leading to margin loss.

SuplaiCerdas was created to mitigate this risk.

Sample material price

What I Learned (and Applied) from the Course

Day 1B: Multi‑Agent Systems & Workflow Patterns

Key concept: Sequential Workflow Pattern – break tasks into a pipeline of specialized agents.

Applied to SuplaiCerdas:

Day 2A: Building Custom Function Tools

Key concept: Custom Function Tools – expose Python functions to agents via the ADK (Agent Development Kit).

Implemented two tools:

def check_component_prices(component_code: str) -> dict:
    """
    Connects to the supplier database and returns a dict:
    { "Supplier_A": price_a, "Supplier_B": price_b, ... }
    """
    # (implementation omitted for brevity)
    pass

def calculate_price_change(current_price: float, last_month_price: float) -> tuple:
    """
    Returns (percentage_change, risk_level)
    Risk levels:
        HIGH   > 10%
        MEDIUM 5–10%
        LOW    ≤ 5%
    """
    # (implementation omitted for brevity)
    pass

Risk level definitions

Conclusion

The AI‑agents intensive course equipped me with the knowledge to design a sequential multi‑agent workflow, build custom tool integrations, and implement systematic evaluation. By applying these concepts, SuplaiCerdas can proactively monitor component price fluctuations, assess risk, and provide actionable recommendations—helping IoT manufacturers protect their profit margins throughout the lengthy tender‑to‑production cycle.

For the full source code and additional documentation, visit my GitHub profile: @leonhardsuharjo.