Intelligent Handloom Commerce & Customer Success Hub

About the Customer

Amounee is a handloom and handicraft company and D2C sustainable fashion brand that works directly with artisans from 75+ clusters across India, helping them reach customers globally. The platform offers handwoven sarees, dupattas, shawls, stoles, and curated craft products, with each piece handmade and naturally unique.

Customer Challenge

Amounee receives around 130 customer enquiries daily, all handled by a single support representative. Approximately 78% of these enquiries come via phone calls from women aged 35–60 who prefer speaking over chat, and most questions are repetitive-shipping timelines for specific clusters, product care for natural dyes and handloom weaves, COD eligibility, returns and exchanges, wholesale pricing, and authenticity verification.

High call volume meant that about 15% of enquiries (around 20 per day) went unanswered. With a 20% enquiry-to-order conversion rate, this translated to roughly 4 lost orders daily, or around 88 missed sales per month. Hiring more staff would have reduced pressure, but founder Megha wanted to avoid adding fixed headcount and management overhead, raising a key question: how could Amounee handle more enquiries without hiring more people?

Without addressing this bottleneck, Amounee would continue losing dozens of potential orders each month, risk damage to its brand from missed calls, and cap growth at what a single person could handle on the phones.

Partner Solution

ISAGA partnered with AWS to build and deploy an AI-powered voice support platform that solved Amounee's enquiry handling bottleneck. During the pre-implementation phase, ISAGA conducted discovery sessions to understand Amounee's call patterns, common questions, and business rules, then designed a solution architecture that integrated Exotel telephony with AWS AI services. When customers call Amounee's support number, they now connect to an AI agent that answers in natural Hindi or English, handling common questions about orders, products, and policies, while seamlessly transferring complex cases to human agents with full context.

The solution leverages several AWS services, each delivering specific capabilities. Amazon Bedrock with Claude Sonnet serves as the conversational reasoning engine, configured with custom prompts that embody Amounee's warm, artisan-focused brand voice. ISAGA implemented multi-turn conversation handling so the AI can manage complex policy discussions-for example, determining COD eligibility by checking customer history and applying partial COD rules. Amazon Nova Sonic 2 is deployed for real-time speech-to-speech processing, configured with native Hindi and English support and Indian accent recognition to handle code-mixed conversations common in Amounee's customer base.

Amazon ECS (Elastic Container Service) hosts the LiveKit-based WebRTC infrastructure and agent management workers, deployed across multiple Availability Zones for high availability. ISAGA configured auto-scaling policies to handle peak call volumes during festival seasons. Amazon OpenSearch powers the retrieval-augmented generation (RAG) pipeline, with ISAGA building and maintaining a vector database containing product FAQs, handloom care guides, weaving technique explanations, and policy documentation. The RAG system enables sub-100ms retrieval of relevant information during calls.

Exotel provides carrier-grade inbound call routing and SIP connectivity into the AWS environment, ensuring reliable telephony infrastructure while AWS services handle AI reasoning, knowledge retrieval, and call context management. Post-implementation, ISAGA provides ongoing managed support including prompt engineering and tuning based on call transcripts, knowledge base updates as Amounee's product catalog and policies evolve, performance monitoring and optimization, and regular reviews to identify new automation opportunities as call patterns change.

Results and Benefits

Before the AI voice agent, Amounee missed roughly 15% of daily enquiries-around 20 missed calls per day. With a 20% conversion rate and an average deal size of about ₹15,000, this translated to approximately ₹60,000 in lost revenue per day. By answering these previously missed calls, the AI voice agent helped recover an estimated ₹1.32 million in monthly revenue, driving an ~18% increase in overall revenue.

The AI agent now handles about 22% of all enquiries end-to-end without human intervention-roughly 29 enquiries per day-freeing up around two hours daily for the support representative. That time is now focused on higher-value conversations and closing new deals, rather than answering the same routine questions.

Call answer rate has increased from 85% to 100%, average response time has dropped from 2–3 minutes to around 4 seconds, and Amounee has achieved 24/7 availability without extending human working hours. With a scalable, voice-first support layer in place, the company can now invest more aggressively in marketing and growth without proportionally increasing support headcount.