There are 37 days left until the 11.11 sale and 53 days until Black Friday on November 27, 2026 (counting from October 5). In that time, an e-commerce store or marketplace seller can still launch AI-powered tools that take pressure off the team during peak days: answers to common questions about delivery and promotion terms, prep for promotional catalog pages, checks of sale prices against margin, inventory forecasts, and handling reviews and returns after the peak. New autonomous agents with permission to change prices or orders should not be launched a week before the sale: during peak days, the cost of a mistake is highest.
Prepared by AIrassvet October 5, 2026 for e-commerce store owners, e-commerce leaders, and marketplace sellers. Marketplace sale dates change from year to year — check the current promotion calendar in the seller portal. AIrassvet is not affiliated with the platforms mentioned; the section at the end describes the publisher’s services.
Contents
- 2026 Sale Season Calendar
- What Happens to Your Store During Peak Days
- Six Scenarios You Can Still Launch Before the Peak
- Support: AI Assistant for Promotion Rules
- Content: Product Pages, Descriptions, and Email Campaigns
- Pricing and Inventory: Checks, Not Autopilot
- After the Peak: Reviews, Questions, and Returns
- Weekly Plan
- What Not to Launch Before the Peak
- Frequently Asked Questions
- Primary Sources and Verification Method
2026 Sale Season Calendar
| Event | Date | Comment |
|---|---|---|
| 11.11 Sale | November 11, 2026, Wednesday | Marketplaces usually extend the promotion for several days before and after — check the timing in the seller portal |
| Black Friday | November 27, 2026 | Friday after the fourth Thursday in November (November 26) |
| Cyber Monday | November 30, 2026 | Monday after Black Friday |
| Holiday Sales | December 2026 | The next peak, when returns and questions shift over |
The dates for Black Friday and Cyber Monday are calculated from the calendar. Each platform publishes its own specific Ozon, Wildberries, and Yandex Market promo periods, application deadlines, and discount requirements.
Participation in marketplace promotions is voluntary. In February 2025, the Federal Antimonopoly Service (FAS) required Ozon and Wildberries to ensure a voluntary, transparent, and predictable mechanism for seller participation in sales by February 28, 2025, after products were included in promotions without sellers’ consent (Habr, February 5, 2025). That means the decision to participate and how deep to discount is yours, and it should be made based on calculations, not inertia.
What Happens to Your Store During Peak Days
A sale puts pressure not only on the warehouse and the site. Everything rises at once:
- customer questions — about delivery times, stock availability, promotion terms, promo codes, and order status;
- content workload — promotional assortments need to be produced fast: product pages, descriptions, banners, and email campaigns;
- pricing errors — a discount applied to the wrong item or below cost creates losses on every order during peak hours;
- stockouts and overstock — popular items sell out, while items purchased for the promotion are left over;
- reviews, product questions, and returns — the wave arrives one to two weeks after the peak and overlaps with New Year prep.
This is exactly where AI is useful: it handles high-volume text and data work, where people become the bottleneck.
Six Scenarios You Can Still Launch Before the Peak
| Scenario | What AI Does | Launch Timeline | Risk of Error |
|---|---|---|---|
| Answers to Common Questions | Replies about delivery, payment, and promotion terms using the store’s rules database; escalates complex cases to an operator | 3–4 weeks | Medium: response database needs control |
| Promotional Product Pages | Prepares descriptions, specifications, and copy for banners and email campaigns; a person reviews them | 2–3 weeks | Low: published by a person |
| Promotion Price Check | Compares the promotional price against cost, commission, logistics, and minimum margin; flags risky items | 2–3 weeks | Low: AI only flags items |
| Inventory Forecasting | Estimates demand for promotional items based on the history of past sales events | 4–6 weeks | Medium: depends on data quality |
| Replies to Reviews and Product Questions | Drafts responses, identifies negative feedback and recurring complaints | 1–2 weeks | Low with moderation |
| Sorting Returns and Claims | Classifies requests, gathers data, and prepares a resolution for an employee | 2–3 weeks | Medium: human decision required |
Timelines are an editorial estimate for a task where the data is available and integration with core systems is possible. If the data is missing or scattered across sources, the timeline gets longer.
Support: AI assistant for promotion rules
The fastest impact during a sale comes from an AI assistant that answers common questions using the store’s rules database — about delivery times and costs, promotion terms, payment, returns, and order status. These questions account for a high-volume stream that grows many times over at peak, and the answers are already written into the rules.
What matters for peak days:
- Answer database — only from approved rules. The assistant must answer based on the document with promotion and delivery terms, not from the model’s “memory.” This is RAG technology; for details on its accuracy, see the article “RAG Accuracy and Hallucinations”.
- Promotion terms are uploaded in advance and can be updated in one click. If the platform or store changes the terms during the promotion, the assistant needs to know immediately.
- Order status — through integration, not guesswork. Without access to the order system, the assistant should honestly route that question to a human agent.
- Handoff to a person — based on keywords (complaint, claim, refund), emotional tone, and the customer’s request.
- Load testing before the peak. Check the model provider’s API limits and the assistant’s behavior under simultaneous conversations. Peak time is not when you want to discover the limits.
More on scenarios and architecture in the articles “AI for Customer Support” and “AI Chatbot for Your Website”.
Content: product cards, descriptions, and campaigns
Promotional assortments often include items that used to sell poorly and were put together in a rushed way. AI speeds up preparation:
- descriptions and specifications based on supplier data and the internal catalog;
- headline and banner text options for different segments;
- email and push notification copy that reflects the promotion terms;
- product card checks for errors: blank fields, conflicting specs, incorrect units.
A person should publish. AI prepares the draft, and the manager reviews and publishes it. This is especially important for specifications: an error in size or composition leads to returns, which tend to arrive right after the peak. Marketplace best practices are covered in the article “Product Pages for Ozon and Wildberries with AI”.
For email campaigns, remember the rules for promotional messages: a send without the recipient’s consent can cost more than it brings in. More on this in the article on fines for AI spam.
Prices and inventory: verification, not autopilot
During a sale, AI is more useful as a price checker than as a price setter. A simple, reliable scenario is checking promotional prices before launch:
- for each item, margin after the discount is calculated taking into account cost of goods, marketplace commission, logistics, and storage;
- items where the margin is below the set threshold or negative are flagged for manager review;
- errors are checked separately: a discount on a non-promotional item, a price below purchase cost, and a mismatch between the website price and the marketplace price.
The check itself is just arithmetic, and it can be done with formulas. AI adds value where the data is scattered: matching items from different exports, parsing supplier price lists in arbitrary formats, and explaining to the manager why an item was flagged.
Demand forecasting for promotional items requires a history of past sales events and time for validation. If you start in October, the model will have time to provide an estimate to support a purchasing decision at least for Black Friday. Details in the articles “Demand Forecasting with AI” and “Dynamic Pricing with AI”.
After the peak: reviews, questions, and returns
One to two weeks after the sale, the second wave arrives: reviews, product questions, returns, and claims. It’s convenient to prepare for it in advance, while there is still time before the peak.
- Reviews: AI drafts responses, highlights negative feedback, and groups repeated complaints — for example, “runs small” or “wrong color.” That’s a signal to fix the product page before the next promotion. More in the article “AI for Review Responses”.
- Returns and claims: AI classifies the request, gathers order data, checks deadlines, and prepares a resolution for the employee. The refund decision remains with a person. Scenarios are covered in the material “AI for Handling Claims and Returns”.
- Catalog search: if shoppers can’t find products during the sale, the issue may be search. More on this in the article “Smart Search for E-Commerce Stores”.
Week-by-week plan
The plan is based on a start date of October 5, 2026, and preparation for Black Friday. If you are targeting 11.11, shorten each stage.
- Week 1 (October 5–11). Choose scenarios and data. Choose two or three scenarios from the table, gather delivery and promotion rules, product exports, history from past sales events, and examples of customer questions.
- Week 2 (October 12–18). Prototype. Rule-based support assistant, product card generation for part of the promo assortment, price-check script.
- Week 3 (October 19–25). Test on real questions. Run the assistant on questions from past promotions, measure the share of correct answers and handoffs to agents, and fix the rules database.
- Week 4 (October 26–November 1). Integration and load testing. Connect the order system, test API limits, configure human handoff and logging.
- Week 5 (November 2–8). Go live in production before 11.11. The assistant works with live shoppers under supervision, and the product cards and prices are verified.
- Weeks 6–7 (November 9–22). 11.11 peak and review. Review conversations handed off to a human and unanswered questions every day; keep expanding the rules database.
- Week 8 (November 23–29). Black Friday. Any changes to the assistant should be limited to fixes in the answers database, with no new features.
- December. Second wave. Reviews, returns, and preparation for holiday sales.
What Not to Launch Before Peak Season
- Autonomous price changes. An agent that changes prices on its own during a peak day can create losses in an hour that the entire promotion will not make up for.
- Autonomous refund decisions. A mistake in favor of or against the customer during a sales peak gets multiplied across hundreds of orders.
- New integrations one week before the promotion. Anything connected to orders, payments, and inventory needs to be tested in advance.
- An assistant without human handoff. A shopper who can’t reach a support agent during a sale will go to a competitor and leave a review.
Rule for peak weeks: One to two weeks before the sale—fixes only, no new features. Anything that wasn’t tested in time gets pushed to December.
Frequently Asked Questions
What can be automated with AI for Black Friday?
You can automate answers to common questions about shipping and promotion terms, prepare product pages and copy for promotional items, check promo prices against margin, forecast demand for discounted products, draft replies to reviews, and sort returns. Pricing and refund decisions are better left to people.
When is Black Friday in 2026?
Black Friday in 2026 falls on November 27, and Cyber Monday is on November 30. The 11.11 sale takes place on November 11. Marketplace promotion periods vary and are published by the platforms themselves in the seller dashboard.
Will I have time to launch an AI assistant before the sale?
If shipping and promotion rules are already compiled and the order system supports integration, a basic store-rules assistant can be launched in three to four weeks, including testing against questions from past promotions. Without testing and human handoff, it should not be launched at peak time.
Is it mandatory to participate in marketplace sales?
No. In 2025, Russia's Federal Antimonopoly Service required Ozon and Wildberries to provide a voluntary and transparent mechanism for seller participation in promotions. The decision to participate and the discount level should be based on a margin calculation for each item.
Can AI be trusted to change prices during a sale?
We do not recommend doing this before peak season without extensive testing. It is safer to use AI as a monitor: it flags low-margin or negative-margin items and pricing errors, while a manager makes the decision.
What should be done with reviews and returns after the sale?
Prepare second-wave scenarios in advance: draft replies to reviews with moderation, group recurring complaints to update product pages, and classify returns with a recommendation prepared for the employee.
Primary Sources and Verification Method
Verified on October 5, 2026:
- Habr, February 5, 2025 — the FAS requirement for Ozon and Wildberries to ensure a voluntary, transparent, and predictable mechanism for seller participation in promotions by February 28, 2025.
The dates for 11.11, Black Friday, and Cyber Monday in 2026, as well as the number of days until them, were calculated using the calendar as of October 5, 2026. The table of scenarios, launch timelines, weekly plan, and list of what should not be launched before peak is an editorial analysis by AIrassvet. Timelines depend on the condition of the data and systems of a specific store and are not a promise.
How AIrassvet Helps Prepare a Store for Sales
If you want to lighten the load on support and the content team for the sale season without risking prices and orders, the task can be broken down step by step. On the AIrassvet website, which also uses the name 'AI Studio by Vladimir Lomtev,' the following directions are described as applicable to it:
- An AI assistant for responding to customers, integrated with communication channels and the order management system, with complex questions handed off to a human agent.
- A RAG system for shipping, payment, and promotion rules so the assistant answers only from approved documents.
- Generative preparation of product pages, descriptions, and campaign copy for promotional assortments with human review.
- ML demand forecasting and margin-based price checks for purchasing and promotion participation decisions.
The first step is to choose one process, describe its current metrics, data sources, constraints, and the criterion by which you will consider the launch successful.
Bottom line: get the essentials done and avoid risk at peak
There is still time before 11.11 and Black Friday to launch AI tools that reduce heavy workloads: answers to common questions about store rules, product page preparation, promo price checks, and scenarios for the second wave of reviews and returns.
Anything that changes prices, orders, and money autonomously should be left for a calm period. One to two weeks before peak—fixes only. That way, AI helps you get through the sale instead of becoming its biggest risk.