Black Friday Customer Service Rarely Breaks on Volume. It Breaks on Repetition.


Hundreds of identical questions arrive in the same four days. This 10-step checklist shows what to automate, what to keep human, and how to score your team’s readiness before the peak.

Peak season doesn’t create new support problems. It multiplies the ones you already have, and it does it in the four busiest days of your year.
Black Friday customer service rarely fails because of one big problem. It fails because of hundreds of small, identical questions arriving at once: where is my order, does this code work, can I return this after Christmas.
Each question is easy. Together, they bury your team right when every minute counts.
And the volume is concentrated in very few days. According to Adobe Analytics data reported by Talk Business, US shoppers spent $44.2 billion online during Cyber Week 2025, and phones accounted for 56.4% of online transactions over the season. That means more shoppers asking questions from a small screen, in the middle of a purchase, expecting an answer in seconds.
Teams are already responding with automation. Salesforce’s Cyber Week 2025 data shows agentic customer service conversations grew 55% week over week.
The question is no longer whether to automate. It’s what to automate first.
Before going further, the obvious disclosure: Landbot builds AI agents for customer service, on websites and on WhatsApp. The checklist below works with whatever tool you use. Where we mention Landbot, it’s to show what a step looks like in practice.
During Black Week, deflection is the number that decides whether your team spends Friday on complaints that need a person or on the fortieth “where is my order?” of the hour.
But deflection only counts when the question is actually resolved. An agent that loops a frustrated shopper, or hides the way to a person, doesn’t reduce tickets. It delays them and makes them angrier.
Deflection that hides the human isn’t deflection. It’s abandonment.
Before you build anything, measure your exposure. These are the four questions to take into your next team meeting, scored 0 to 10 each (0 = this breaks on Friday, 10 = no change from a normal day).
You can’t hire your way out of a low Coverage score. You can only write your way out.
Export last year’s tickets from late November and early December and tag the top ten reasons for contact. This list decides everything else: which answers to write, which agents to build, and where you need people.
Which products are included, whether codes stack, when each offer starts and ends. Put it in one document that your site, your agent, and your team all read from. Most Black Week confusion comes from three versions of the same rule.

Shipping cut-offs for Christmas delivery, the extended return window, and what happens with gifts. Publish them on one page your agent can answer from, and link it in order confirmation emails.
Order status (WISMO, in support jargon) is usually the biggest ticket category of the season. An agent that asks for the order number and email, shows the status, and opens a case only when the order is late removes a large share of repetitive work.
With Landbot, you connect your order data through native integrations and webhooks and configure the flow yourself, without waiting on IT.

Decide which conversations go to a person, who covers which hours, and what information arrives with the handoff. A handoff that passes the order number and the question means your team never has to ask twice.

Your website handles discovery and self-service; WhatsApp handles follow-ups, cart reminders, and shoppers who arrived from ads. Use the same rules and the same agent logic on both, so answers never contradict each other. Our WhatsApp business automation guide covers the channel setup.
The best ticket is the one that never arrives. Send shipping updates, delay notices, and “your offer ends tonight” reminders before shoppers ask. On WhatsApp, get your message templates approved well before the sale starts.

Resolution rate, escalation rate, and first response time. If escalations spike on one topic, the fix is usually a missing answer in your knowledge base, not more people.

Holiday returns arrive in December and January. A guided returns flow that checks eligibility, offers an exchange first, and records the reason keeps that wave manageable and saves part of the sales.

Compare this year’s numbers against last November: tickets per order, resolution rate, CSAT. Most of the agents you built (order status, returns, sizing) stay useful all year. Only the promotion rules need to change.
Black Friday customer service rewards preparation, not heroics.
Start with the top three ticket reasons from last year, automate them, test them with real questions, and keep your team for the conversations that need a person. You can start from our AI agent templates or build your own with the AI agent builder. Try Landbot free and have your first agent live before the rush.
Six weeks is comfortable; two weeks is the minimum. Leave the final week for testing with real questions from last year, not for building.
WISMO stands for “Where Is My Order?”, the most common support question in e-commerce. In peak season it often tops the ticket list, which makes it the first flow worth automating.
Whatever topped last year’s ticket list. For most stores that’s order status, followed by promotion questions and returns.
Not if the handoff is visible. Shoppers accept automated answers when they’re fast, correct, and one tap away from a person.
Often fewer than you think once repetitive questions are automated. Plan staffing around complaints, exceptions, and high-value orders instead of total ticket volume.
Many retailers extend returns for November purchases until mid or late January so gifts can be returned after the holidays. Check your legal minimum first (in the EU, online buyers have a 14-day right of withdrawal), then state the extended window clearly at checkout and in your agent.
Both, with one brain. Use the website for discovery and self-service, WhatsApp for follow-ups and proactive updates, and the same rules on each channel.
Yes. In Landbot you build the flows, connect your order data through integrations or webhooks, and edit answers in production yourself. A first order-status agent usually takes days, not weeks.
Track tickets per order, resolution rate, escalation rate, and CSAT, and compare them with last November.