← Back to case studies

    E-commerce / Custom Manufacturing

    Custom Cabinetry E-commerce Brand

    A custom AI-powered operations layer that closed the gaps Magento couldn't, supporting a custom cabinetry e-commerce business with $10K+ orders.

    AI/OCRCustom ReactMagento APIsZoho CRMRetellPallet packing algorithmParallel API orchestration

    Impact

    Annualized impact

    0

    Minutes saved per shipping quote (avg of 20 to 30)

    Time to build a shipping estimate

    Manual, frequently inaccurate25min
    Automated, 95% accurate1min

    Per quote, repeated dozens of times daily

    01

    18+

    Routes on the voice agent (most complex agent we've built)

    02

    20-30 min

    Saved per shipping quote

    03

    95%

    Accuracy on automated pallet packing estimates

    04

    $10K+

    Average order value supported

    01 // The challenge

    The challenge.

    Bridge the gaps in Magento and Zoho that were costing this brand accuracy, speed, and customer service capacity on high-value custom orders.

    • 01Magento couldn't swap colors on a configured product build without manually re-adding every SKU
    • 02Shipping quotes for bulky custom orders took 20-30 minutes per request and were often inaccurate
    • 03Customer service reps had to dig across disconnected Zoho objects to get a full picture of any one customer
    • 04Manual invoice review before manufacturer handoff was error-prone and expensive when missed

    Why off-the-shelf wouldn't work

    This is a $10K+ AOV custom-build e-commerce business with a CS team running parallel to a designer team, all on top of Magento and Zoho. No off-the-shelf tool sits in that gap. We had to build it.

    02 // What we built

    What we built.

    01

    Cart configuration agent

    AI agent that swaps colors, configurations, and SKUs on custom builds without manual re-entry. Closes the gap in Magento.

    02

    Shipping estimator

    AI plus pallet-packing algorithm that estimates shipping weight and pallet configuration with 95% accuracy in seconds, not 20-30 minutes.

    03

    Invoice OCR system

    Compares incoming order invoices against manufacturer purchase orders to catch discrepancies before they become costly errors.

    04

    Unified customer view

    One-screen customer view that aggregates Zoho objects, order history, and rep activity by punching in a phone number, email, or order ID.

    05

    AI voice agent

    18+ routes including assembly help, designer scheduling, returns and damages, sales and specials. Handles tier-one volume so the team can focus on complex builds.

    06

    First-layer chatbot

    Triages chats, hands off to live team, and handles after-hours volume from West Coast customers when the East Coast team is offline.

    03 // Technical implementation

    Architecture overview.

    A custom React tooling layer that wraps Magento and Zoho through their APIs, with parallel API orchestration powering the unified customer view and AI/OCR powering invoice review and the shipping estimator. The voice agent and chatbot run on Retell with custom tool calls into the same data layer, so customer-facing AI and internal AI share a single source of truth.

    technical-highlights.md

    • → Magento API orchestration for live cart manipulation
    • → Parallel API requests across Zoho objects for unified customer view
    • → AI summarization for real-time customer profile context
    • → OCR-based invoice comparison with manufacturer POs
    • → Pallet packing algorithm for shipping estimation
    • → Retell voice agent with 18+ conversational routes

    04 // What this demonstrates

    Custom internal toolingAI/OCR document processingAlgorithm designE-commerce platform integrationComplex voice agent routingParallel API orchestration

    In their own register

    "
    When Magento couldn't, our custom layer did.
    "
    20-30 minutes per shipping quote, gone.
    "
    18 conversational routes, one voice agent, full tier-one coverage.
    "
    One screen, every customer, every detail.

    Ready to architect your next system?