Multi-Agent Ops Automation for FMCG Brand

Nov 29, 2025

About the company

Our Client is a fast-growing commerce brand operating across Shopify, Amazon, and direct channels. The team manages high-volume order fulfilment, catalogue operations, refunds, and paid acquisition. Because each function runs on different systems including Shiprocket, Razorpay, Odoo, Shopify, Amazon, and Meta Ads, Saarvay prioritized speed, accuracy, and consistency in day-to-day operations to keep margins healthy while scaling.

The problem

Saarvay's operations relied on repetitive, manual workflows spread across multiple platforms, creating bottlenecks and error risk.

• Shiprocket: Manual carrier selection on cost and rating, invoice generation, and label printing for each order.
• Odoo: Repetitive purchase order creation, inventory updates, and material count adjustments prone to mistakes.
• Shopify/Amazon: Heavy product listing work with large forms, image uploads, and frequent AI-based image enhancement before publishing.
• Refunds: Case-by-case email processing with checks for time since purchase, reason, and amount.
• Meta Ads: Manual KPI tracking and budget adjustments to shift spend from losing to winning campaigns, requiring constant monitoring.

The result was significant analyst time spent on routine clicks, higher risk of errors, and the need for dedicated staff just to keep daily operations moving.

The solution

We deployed a fleet of specialized automation agents that handle each workflow end to end, with human-in-the-loop controls where judgment is required.

Shiprocket Agent

• Chooses carriers using rules for cost, rating, and product type.
• Generates invoices and prints labels automatically via the interface.
• Surfaces exceptions such as no serviceability for quick human review.

Odoo Agent

• Automates purchase order creation and material count updates.
• Integrates with WhatsApp to send PO alerts, low-stock notifications, and accept structured updates through chat.

Shopify/Amazon Agent

• Automates new product creation with form filling and catalogue updates.
• Runs AI-based image enhancement before upload.
• Includes human-in-loop review gates for titles, attributes, and images before publishing.

Refunds Agent

• Verifies each request against policy rules including time window, reason, order value, and payment method.
• Approves or rejects, triggers notification emails, and logs outcomes to the database.

Meta Ads Agent

• Tracks KPI performance such as CTR, CPC, CPA/ROAS and classifies winners/losers.
• Adjusts budgets based on predefined thresholds and pauses underperformers automatically.
• Flags anomalies and sends summaries to the team.

Success metrics

• 50% reduction in operations headcount needed for day-to-day processing.
• Faster Shiprocket turnaround from order ready to label printed and dispatched.
• Fewer human errors across inventory, listings, and refunds due to standardized automation steps.
• Lower learning curve for complex tools like Odoo thanks to guided chat flows and automated updates.
• Higher overall operational efficiency, enabling the team to focus on growth initiatives.

Business impact: Leaner operations, faster fulfilment cycles, and improved data quality across the commerce stack.