Black Friday falls on Friday, November 27 this year, with Cyber Monday on November 30. Thanksgiving is the day before. Sales often start well ahead of those dates, which is exactly why the next week is when shoppers make fast decisions based on star ratings and a crossed-out price. Both of those signals can be manipulated, and a chatbot is surprisingly good at poking holes in them if you give it the right job. (Sources: SoFi, Woman & Home)
This is a practical workflow, not a promise. AI cannot see a seller's server logs, and it cannot prove a review is fake. What it can do is read 50 or 100 reviews faster than you can, notice patterns a human skims past, and force you to compare the claimed discount against the price history. Used that way it can shorten a long research spiral to something closer to 10 minutes.
Quick summary: Do five things before buying any big-ticket Black Friday item. 1) Check the price history, not the percent off. 2) Paste the 3-star and 1-star reviews into an AI and ask for recurring complaints. 3) Ask the AI to flag suspicious review patterns, as hints only. 4) Look for the same specs under a cheaper or older model name. 5) Verify warranty and return terms in writing. Never paste payment details or account logins into a chatbot.
Why reviews and discounts need a second look
The Federal Trade Commission thought the problem was big enough to write a rule. Its final rule, announced in August 2024, bans fake or false consumer reviews and testimonials, including reviews that claim to come from someone who does not exist, which the agency said covers AI-generated reviews. It also prohibits paying for reviews conditioned on a particular sentiment, review sites a company controls while posing as independent, and using threats to suppress negative reviews. The rule took effect on October 21, 2024, and lets the FTC seek civil penalties against knowing violators. (Sources: FTC, Alston & Bird)
A rule is not the same as a clean marketplace. It applies to people who knew or should have known a review was fake, and enforcement takes time. Treat the law as a reason bad actors have more to lose, not as a reason to trust every five-star pile on a product page.
Step 1: Check the price history before anything else
A chatbot without live data cannot tell you what a product cost last month, so use a price tracker for this step. For Amazon, tools such as CamelCamelCamel and Keepa show a price chart if you paste in a product link, and both let you set an alert at your target price. Guides that cover them advise looking at the 12-month low and the typical price instead of the percentage off, because the "original" price in a sale can be a number the item rarely sold at. For other retailers, Google's price insights on shopping results can help, though coverage is less complete. (Sources: Trusted Reviews, SlashGear, Android Authority)
Where AI helps: once you have the numbers, give them to the chatbot and ask it to do the arithmetic and the skepticism. This prompt works well:
Prompt: "Here is a product's advertised sale price, its advertised 'regular' price, its 12-month low and its typical price from a price tracker: [paste numbers]. Calculate the real discount against the typical price and against the 12-month low. Tell me whether this is a good deal, an average deal or a marketing discount. Show your math. If any number is missing, say so instead of guessing."
Step 2: Read the 3-star and 1-star reviews, not the 5-star ones
Padded reviews tend to skew toward five stars, and five-star reviews are also often the least informative. Copy 30 to 60 of the lowest and middle reviews into a chat, break the paste into batches if the tool has a length limit, and ask for a summary of repeated problems. A recurring complaint that appears in different words, such as battery dying in the second month or a hinge cracking, is far more reliable than any single story.
Prompt: "Below are customer reviews of a product. Group the complaints into themes. For each theme give a count of how many reviews mention it, the best short quote, and whether the problem appears early (first month) or late. Ignore reviews that only complain about shipping. Do not invent reviews that are not in the text."
Step 3: Use AI to flag review patterns, then verify them yourself
Ask the chatbot to look for hints of padding in a batch of reviews. Useful signals include many reviews posted within a few days, nearly identical phrasing, praise that never mentions a specific feature, reviews about a different product variant, and reviewers whose only activity is a string of glowing posts. Be honest about what this proves. A pattern is a reason to be suspicious, not a verdict, and AI can wrongly flag real enthusiastic buyers. Real products often get review bursts after a promotion or a viral video.
- Sort by newest. A product with a thousand reviews that all landed in one month deserves a closer look.
- Read the verified-purchase and unverified reviews separately where the site shows them. Different tone between the two is a hint.
- Look for reviews that mention a different product, or describe features the item does not have.
- Check whether negative reviews are answered with scripts or deleted. The FTC rule specifically covers review suppression by businesses.
- Search the product name plus the word "problem" or "return" on a forum site to find reviews outside the seller's page.
Step 4: Hunt for the same product under a different name
Retailers sometimes sell a model variant built just for a sale event, with a slightly different part number and often a lower spec. Ask the chatbot to compare the model number on the sale page against the manufacturer's current product line, and paste in the spec sheets from both. Models fresh from last year are often the real bargain, and the AI can tell you what actually changed. It must be told to use only the pages you paste, or it may fill the gaps from memory, which for a fast-moving product category is stale.
Step 5: Read the return and warranty terms
A great price loses its shine if the return window is short or restocking fees apply. Paste the return policy and ask for the deadlines, fees, exceptions and who pays shipping, then confirm them on the retailer's page yourself. Take a screenshot of the policy at checkout time, because holiday return windows can differ from normal ones.
How the five checks compare
| Check | Best tool | What it tells you | Limit |
|---|---|---|---|
| Price history | Price tracker (Keepa, CamelCamelCamel) | Whether the discount is real | Mostly Amazon; other stores are patchier |
| Review themes | Chatbot with pasted reviews | Recurring defects in plain language | Only as good as the sample you paste |
| Padding patterns | Chatbot plus your own eyes | Reasons to be suspicious | Cannot prove a review is fake |
| Model comparison | Chatbot with spec sheets | Whether a cheaper variant is weaker | Needs the real manufacturer pages |
| Return terms | Chatbot plus the retailer page | Deadlines and fees | Policies can change at checkout |
A 10-minute routine you can reuse
- Minutes 1 to 2: paste the product link into a price tracker and note the 12-month low and typical price.
- Minutes 3 to 5: copy the lowest-rated reviews into a chat and ask for repeating themes.
- Minutes 6 to 7: ask for padding hints, then sort by newest to confirm or reject them.
- Minutes 8 to 9: compare the model number to the manufacturer's page.
- Minute 10: screenshot the return policy and set a price alert if the deal is not good enough yet.
Where AI gets this wrong
Three failure modes show up repeatedly. First, a chatbot can sound certain about a price it never saw. Never accept a number without a source. Second, it can invent a review quote if you ask for "examples" without telling it to quote only from your paste. Third, it tends to agree with your hunch. If you say, "I think these reviews are fake," it may go along with you. Ask it for the strongest case that the reviews are genuine, then compare.
Privacy matters too. Do not paste card numbers, order confirmations with your address, or account logins into any chatbot. Product pages and public reviews are fine. Anything tied to your identity does not need to be there for this job.
If you want a place to run these steps, LumiChats lets you work in one chat and switch between AI models without juggling separate subscriptions, which helps when you want a second opinion on a review summary or a spec comparison. Whichever tool you use, the habit is the same: bring your own evidence and ask for the math.
01Can AI tell if a review is fake?
No. It can highlight patterns that deserve suspicion, such as repeated phrasing or bursts of posts, but it cannot verify who wrote a review. Treat its output as a lead and confirm with your own check.
02Are fake reviews illegal in the US?
The FTC's final rule on reviews and testimonials, effective October 21, 2024, bans fake or false reviews and lets the agency seek civil penalties against knowing violators. Enforcement still depends on the FTC finding and pursuing cases.
03Is Black Friday really the cheapest time to buy?
Not always. The date is Friday, November 27, 2026, but price history is what answers the question for a specific item. Some products hit their low in October or after the holidays.
04Which price trackers work?
CamelCamelCamel and Keepa are the two most often recommended for Amazon. Both offer alerts. For other retailers coverage is thinner, so compare against the manufacturer's own price too.
05Should I paste my order details into a chatbot?
No. Keep payment details, addresses and logins out. You only need product pages, public reviews and policy text.
