Guide
Black Friday and Christmas 2026: how to prepare customer service with an AI assistant
Short answer
The peak puts the most strain on customer service. Four groups of questions multiply: delivery deadlines, stock, returns and gift shopping. The assistant takes the fact-based questions, and hands over the individual cases and upset customers.
The Black Friday and Christmas season puts the most strain on customer service. Traffic can be scaled: servers, advertising, stock. Human answering cannot. At the peak, more questions arrive, and they are of a different kind: delivery deadlines, stock, returns, gift shopping.
Most of the preparation is written work. The assistant gives as many answers as you write down for it. Below are the four questions that multiply, the preparation of the knowledge base, and what the assistant takes on and what it hands over.
Source: Own platform data, 90 days of use, 2026-07-08 – 2026-10-06
Which questions multiply
At the peak, four groups of questions grow. Delivery deadlines: will it arrive by Christmas, and until when is it worth ordering. Stock: is it still available, and when will it be back. Returns: how, by when, and at whose cost. Gift shopping: wrapping, gift vouchers, a personal message.
Of the four, the delivery deadline is the most pressing. The customer sees the promotion but not the dispatch deadline, and that is why they ask. Stock questions often come together with size: is it available in the size they are looking for. Returns come after Christmas, together with exchanges. In gift shopping, wrapping and gift vouchers are decisive.
Based on our own platform data, 51.8 percent of visitor questions arrive outside business hours (customers in Hungary, CET). A small online store typically receives 15–60 questions a month. The peak adds another layer on top of this narrow band, just when nobody is at the keyboard.
It is worth looking at questions by type, because the logic of the answer also differs.
| Type of question | Who handles it |
|---|---|
| Delivery, fees, promotion terms | The assistant, from the knowledge base |
| Stock, size, specifications | The assistant, from the product data |
| Order status, invoice | The assistant |
| Complaint, damaged product | Your colleague |
| Individual discount, payment dispute | Your colleague |
What to put in the knowledge base in advance
The assistant answers from the company’s own knowledge base. If there is no content for something, it does not make it up: it asks for contact details and hands over to a colleague.
Make the following available in advance: delivery times and fees, the free-delivery threshold, the current promotion terms, the returns process, the warranty and sizing. Also write down the most common questions and good answers to them. If something is not written down, there is no answer to it: only a handoff.
It is worth recording the promotion rules word for word: the exact terms of the promotion, what it does not apply to, and what happens if stock runs out. At the peak, these are the most common questions and misunderstandings.
From mid-December, the delivery deadline is the first question. Put the courier services’ last dispatch dates into the knowledge base and onto the product page. Check this again every year, because it changes from year to year.
Preparation in chronological order
The order of the tasks matters more than their number. The knowledge base decides the rest, so it is worth putting it together first.
Early autumn: the knowledge base. Collect in one place the delivery times and fees, the free-delivery threshold, the returns policy, the warranty and sizing. Write down the most common questions and the good answers. The assistant will not be smarter than this.
Mid-autumn: integration and product data. The assistant is useful when it gives the real price, stock and specifications. It is worth sorting this out months before the peak.
5 November: the order deadline for the assistant to be working before Black Friday. After that, a live test: ask the assistant your own hardest questions. Check the handoff too: does it work, and who receives it.
Mid-November: many retailers already start their promotions at this point. On key pages, it is worth communicating delivery and promotion information with a proactive greeting.
Black Friday and Cyber Monday: at this point you only need to monitor, and route to a human where necessary. Two things are worth measuring: the conversations closed by the assistant, and how long it takes for someone who needs a human to reach one.
December and after Christmas: the delivery deadline is the first question, and then the wave of returns arrives. A step-by-step description of returns is content the assistant will repeat accurately, if you have written it.
What the assistant takes off your hands, and what it hands over
The assistant takes the repetitive, fact-based questions: delivery, fees, promotion terms, stock, order status, invoices, payment methods, the returns process, related products. It answers these from the knowledge base, and the source is shown next to the answer.
What it hands over to a human: complaints, damaged or faulty products, individual discounts, larger bulk quotes, payment disputes, and anything that makes the customer upset. In these cases the assistant hands over with the full history, so that your colleague does not have to ask everything again from the beginning.
The disclosure is built in. The header and the first message state that artificial intelligence is answering, in line with Article 50 of the EU AI Act. The disclosure cannot be switched off. The data stays in the EU, including model calls; retention is 90 days.
When setting it up, two questions matter: how many questions it answers well, and whether it hands over the rest in time.
This is where coverage at night and at the weekend comes in. Half of the questions arrive outside business hours, when nobody is there to answer. At these times the assistant answers from the knowledge base and lists by the morning what was asked.
Timing: when it goes live
Onboarding: live within two working days, if we have access to edit your website or your web developer places the code snippet. An order placed by 5 November goes live before Black Friday. The onboarding fee is €200 excl. VAT, €100 excl. VAT with a 6-month contract and free with a 12-month contract, and the monthly fee is then reduced by 8 or 15 percent respectively. The Chat AI Starter monthly fee is €14.90 / month excl. VAT, with an allowance of 500 answered questions per month.
The demo is free of charge, with no obligation to order; we send the link to the verified demo within 1 working day. When testing, ask your own hardest questions. If it answers badly, something is missing from the knowledge base.
Next step
The knowledge base decides the peak. Request a demo on your own website, or see the prices. If you have a question, write to .
Methodology
This guide is based on our own platform data (90 days, 2026-07-08 – 2026-10-06) and on our experience of onboarding online stores. We did not carry over international benchmark figures that cannot be verified for Hungarian conditions; in their place we describe the facts and the steps of preparation.
Changelog
- 2026-10-15: rewritten from the smartaichat.hu article “Black Friday and Christmas 2026”
Sources
- Own platform data, 90 days of use, 2026-07-08 – 2026-10-06 · own data · smartaiassist.eu/data/after-hours-questions/
- EU AI Act, Article 50 (transparency obligations) · eur-lex.europa.eu
Found a mistake, or do you have more recent data? Write to
and we will review it.
Frequently asked questions
Black Friday and Christmas 2026
Which questions multiply during the season?
Four groups: delivery deadlines, stock, returns and gift shopping. The other questions remain, they are just pushed back.
What can the assistant do, and what can it not?
It takes the fact-based, repetitive questions from the knowledge base. It hands over the individual cases, upset customers and anything with financial consequences to your colleague.
When is it worth starting for the season?
An order placed by 5 November goes live before Black Friday. It is worth putting the knowledge base and the integration in order beforehand.
Is it mandatory to disclose that AI is answering?
Yes. Under Article 50 of the EU AI Act, it must appear in the header and the first message. The disclosure cannot be switched off.
How long does onboarding take?
Live within two working days, if we have access to edit your website or your web developer places the code snippet.
Related pages
See it with your own content.
The automatic preview is ready after you submit, and is deleted after 24 hours.
In the example conversations, the assistant's answers are given by artificial intelligence. Disclosure under Article 50 of the EU AI Act: "Artificial intelligence, not a human".