The challenge
A UAE-based online fashion retailer was processing 18,000 customer conversations per month across WhatsApp and web chat. During Ramadan and Eid sales campaigns, that volume spiked to over 50,000 - almost triple their normal load.
Their 18-person support team was stretched to breaking point. Average response time during peak periods climbed to 4 hours. CSAT dropped 18 points in two consecutive peak seasons. Senior management had tried hiring additional seasonal agents, but quality was inconsistent and ramp-up time was too long to be useful during a 2-week sale.
The majority of their inbound volume was entirely predictable and low-complexity:
- Order status and tracking - "Where is my order?" - accounting for 34% of all contacts
- Return and exchange requests - 22% of contacts
- Size and fit queries - 18% of contacts
- Product availability - 12% of contacts
- Discount code and promotion questions - 8% of contacts
Only 6% of contacts - complaints, complex disputes, and VIP queries - actually required a senior human agent. Yet the entire team was spending 94% of their time on the 94%.
Why they chose Eshal
The retailer evaluated three platforms before selecting Eshal. The deciding factors were Arabic dialect support and setup speed.
Arabic dialect coverage was critical. Their customer base spans UAE nationals, Saudi shoppers, Egyptian expats, and Gulf Arab residents. A chatbot that responded in formal Modern Standard Arabic felt "robotic and off-brand" (their exact feedback from a competitor trial). Eshal's native Gulf and Egyptian Arabic support felt natural to every segment.
Same-day deployment was non-negotiable. They were two weeks from a major campaign launch. Platforms requiring 4–6 weeks of IT integration were disqualified immediately. Eshal's Shopify integration was live in under 4 hours.
The deployment
Implementation ran over a single business day. The Eshal implementation team connected to their Shopify store, configured Arabic and English conversation flows for the top 5 query types, set escalation rules for returns above AED 1,000 and any complaint mentioning a specific keyword set, and ran the first live conversations at 9pm the same evening.
Day 1 live results: 312 conversations handled. 271 resolved by AI (87%). 41 escalated to human agents - all with full conversation context pre-loaded.
"We went live the same day we signed. By the next morning Eshal was handling more conversations than our entire team did in a full shift - in Arabic and English. Our agents came in to a clean queue with only the cases that genuinely needed them."- Head of Customer Experience, UAE Fashion Retailer (anonymised on request)
Results - Month 1
The first full month of deployment coincided with their spring campaign - historically their second highest-volume period of the year.
Conversation volume: 54,200 total contacts handled during the campaign. The previous year, the same period saw 17,800 contacts - a 3× increase driven by aggressive marketing spend. With the old system, this would have required approximately 36 additional seasonal agents. With Eshal, zero additional headcount was required.
Resolution rate: 86.4% of contacts resolved without agent involvement. The remaining 13.6% were escalated with full context - agent handle time on escalated contacts dropped from an average of 9 minutes to 5.5 minutes because the AI had already collected all relevant information.
Response time: Average first response time dropped from 3.8 hours (pre-Eshal peak) to 12 seconds. CSAT during the campaign scored 4.6/5 - a 22-point improvement versus the previous year's peak.
Language breakdown: 52% of conversations in Gulf Arabic, 31% in English, 11% in Egyptian Arabic, 6% in other languages. All language variants handled natively with no configuration changes between regions.
What happened to the team
A natural concern with any automation deployment is impact on the existing team. In this case, the 18-person support team was fully retained and redeployed:
- 12 agents moved to a dedicated VIP and complex-query queue, handling the 13.6% of contacts that genuinely needed human attention - with 75% higher case quality scores than pre-Eshal
- 4 agents transitioned to outbound customer success - proactively contacting repeat buyers and running satisfaction surveys
- 2 agents were promoted to oversee Eshal's knowledge base and quality-review samples of AI conversations weekly
12-month trajectory
By Month 6, the retailer had extended Eshal to their loyalty programme queries, product recommendation conversations, and post-purchase feedback collection. AI resolution rate held steady at 84–87% across all new workflow types. Monthly net saving versus pre-Eshal baseline: AED 127,000 (approximately USD 34,600) - well in excess of the Eshal subscription cost.