Before the November and December peak, UAE retailers should automate the customer questions that arrive in the biggest numbers and need the least judgement: where is my order, when will it arrive, how do I return it, is it back in stock, why is my promo code not working, and what time does the store close. An AI agent on WhatsApp and web chat can answer these from your live order system in English and Arabic, while abandoned cart and back-in-stock flows recover sales on the marketing side, and people keep refunds and complaints. Do it now because White Friday and the Dubai Shopping Festival arrive back to back, enquiries rise exactly when your team is busiest, and an unanswered message during a sale can easily become a lost order.
This guide covers when the peak lands, what spikes, a six-week countdown from mid-October and the guardrails that matter.
When the peak lands this year
White Friday. This is the Middle East's name for the Black Friday season. Amazon.ae's White Friday page says the sale is usually around the last Friday in November, with early access for Prime members. When we published this guide on 10 October 2026, neither Amazon.ae nor noon had announced its 2026 dates. As a guide only, last year's sales started earlier than the Friday itself: Amazon.ae ran its White Friday Sale from 20 to 30 November 2025, and noon ran Yellow Friday over the same dates. Check both again in early November.
Dubai Shopping Festival. Visit Dubai, the city's official tourism site, lists this year's DSF as 4 December 2026 to 17 January 2027. The previous edition ran for 38 days and, according to Gulf News, featured discounts of up to 75% across more than 1,000 brands and 3,500 stores.
Put together, that is close to two months of promotions running from late November into mid-January, on top of a growing online base. A report by EZDubai and Euromonitor International, summarised by Hidubai Focus, valued UAE e-commerce at AED 42.2 billion in 2025, about 15.7% of all retail sales, after 19% compound annual growth since 2020.
What actually spikes, and how to automate it
Customers do not wait for office hours. Zendesk's CX Trends 2026 research found that 74% of consumers now expect 24/7 service, and 85% of CX leaders say customers will drop brands that cannot resolve issues on first contact. Big retailers are already automating for this. In Salesforce's Cyber Week 2025 data, AI agent customer service conversations on its platform grew 55% week over week, and agent actions such as updating delivery addresses and starting returns rose 70%.
Most peak-season messages fall into a handful of types:
| Enquiry | What drives the volume | Automation fix |
|---|---|---|
| Order status | Bigger baskets, more first-time buyers, payment confirmation worries | Agent looks up the order by number or phone and replies with the live status |
| Delivery times | Courier backlogs, promised delivery slots, gifts with a deadline | Agent reads the courier tracking feed and gives the latest estimate, not a guess |
| Returns and exchanges | Sizing on fashion, gifts, impulse buys during deals | Agent explains the policy, checks eligibility and starts the return request |
| Stock availability | Limited sale stock, sizes and colours selling out | Agent checks live inventory and offers a back-in-stock alert |
| Promo codes | Expired codes, minimum spend rules, excluded brands | Agent checks the code against your discount rules and explains why it failed |
| Store hours and locations | Changed opening hours during sales, click and collect | Agent answers from one maintained list of branches, hours and collection points |
None of these needs creativity. They need accurate data, fast, which is where an agent earns its place.
The six-week countdown
Six weeks from mid-October gets a tested first version live before the sales begin; last year, they began on 20 November. Here is the plan we follow.
Week 1 (from 12 October): Audit last year
- Export last November and December's WhatsApp, chat, email and call logs
- Tag the top 30 questions and how many times each was asked
- Note the busiest hours and days, including late nights and weekends
- Agree the first scope: order status, delivery, returns policy, stock, promo codes and hours
Week 2 (from 19 October): Write the source of truth
- Gather returns, delivery and promo policies into one approved document
- Write answers in English and Arabic, checked by a native Arabic speaker
- Set the handover rules: refunds, complaints, damaged goods, payment issues, angry customers
- Decide who on the team receives handovers, and at what hours
Week 3 (from 26 October): Connect the systems
- Connect the agent to Shopify or your order system for order lookups
- Add courier tracking and live inventory
- Connect the team inbox so handovers arrive with a summary
- Log every conversation to the CRM for reporting
Week 4 (from 2 November): Build the marketing flows
- Abandoned cart reminders for customers who have opted in
- Back-in-stock alerts for sold-out sizes and products
- Promo code rules loaded for every planned campaign
- Post-purchase review requests timed after delivery
Week 5 (from 9 November): Test hard
- Run last year's real questions through the agent and review every answer
- Test Gulf dialect, Arabizi and mixed English and Arabic messages
- Try to break it: wrong order numbers, expired codes, refund demands, abuse
- Check that every handover reaches a person within your target time
Week 6 (from 16 November): Soft launch and prepare
- Go live on one channel first and review conversations daily
- Prepare approved messages for courier delays and stock-outs in both languages
- Set the staff rota for handovers through the sale weekends
- Agree one owner who can update answers in minutes when an offer changes
If you start later, cut scope rather than testing time. An agent that answers four questions well beats one that answers twenty badly.
WhatsApp and web chat agents for order and delivery questions
The flow is simple. A customer messages "Where is my order?" on WhatsApp or your website chat. The agent asks for the order number or matches the phone number, looks the order up in Shopify, WooCommerce or your own order system through an n8n workflow, reads the courier status and replies in seconds: shipped, out for delivery, expected tomorrow. If the order is late, it says so plainly and offers a person.
The same agent can change a delivery address before dispatch or resend a tracking link, if you allow those actions. Each runs as a fixed, logged step in the workflow. If you are weighing up which parts need an AI agent and which only need a rule, our explainer on AI agents vs chatbots vs automations walks through the difference. For WhatsApp costs, Arabic handling and Meta's rules, see our WhatsApp AI agents guide.
The marketing side: carts, stock and reviews
Service and marketing share the same data, so build them together.
- Abandoned cart flows. Send a reminder by email or WhatsApp to customers who have opted in, with the item and a link back to the cart. Keep it to one or two messages, and stop as soon as they buy.
- Back-in-stock alerts. When the agent tells someone a size is sold out, it offers to message them when it returns. That turns a dead end into a list of ready buyers.
- Promo code handling. Code complaints often have a simple cause: the code has expired, the basket is below the minimum spend or a brand is excluded. The agent explains which, in one line. It never invents a code; any goodwill code needs a person's approval.
- Post-purchase reviews. Ask for a review a few days after confirmed delivery, not at dispatch. If the customer has an open complaint, hold the request until it is resolved.
Guardrails that protect your brand
Peak season is when a careless agent does the most damage. Set these rules before launch.
- Refunds always go to a person. The agent can explain the policy, collect the order number, reason and photos, and open a ticket. A human decides and processes every refund.
- Complaints escalate immediately. Damaged goods, missing items, payment disputes and angry messages go straight to your team with a summary, so the customer never repeats themselves.
- Arabic and English with equal care. The agent replies in the customer's language, follows them if they switch, and quotes the same prices and policies in both.
- Honest tone during delays. If a courier is behind, the agent says so, gives the latest estimate and apologises once. It never promises a delivery date the system cannot confirm.
- Admit the limits. When the agent does not know, it says so and brings in a person. "Talk to a person" must always work.
After the peak: what to measure and keep
When DSF ends in mid-January, review the season with numbers rather than impressions:
- First response time, in and out of hours
- The share of conversations resolved without a handover
- Handover reasons, ranked, and how quickly a person picked each one up
- The questions the agent could not answer, which become next year's content
- Revenue recovered by abandoned cart and back-in-stock flows
- Customer satisfaction scores, compared with last year's peak
Keep the agent running after the sales end. Order status and returns questions do not stop in January, and this season's conversations make the agent better for the next campaign.
Start before November
A focused agent with order lookups, handover rules and bilingual testing takes about four to six weeks to do properly, and the integrations and testing are the parts you cannot rush. Starting in October means you go into the sales with a tested system rather than a rushed one.
We are marketers first, so we start from your customer journey and campaign calendar, then build the automation around them. If you want this live before the November sales, book a call with us and bring last year's most common customer questions.
