All use cases

Analyze a month of reviews: what people ask for most

Buyers keep asking for the same small changes, one review at a time. The agent pulls those asks from Ozon and Wildberries into a ranked list with quotes.

Easy · 15 min · once a month

What you get

Customer requests, March — 214 reviews
1. Lid latch — 34 mentions
   "the lid slips off when carried, I hold it by hand"
2. Sizing in the description — 19 mentions
   "ordered M, got a kids size, no size chart anywhere"
3. Spare filters — 11 mentions
One-off asks (1-2 mentions) go on a separate sheet.

A sample on made-up data — your numbers will be your own.

Who it fits

  • Planning next month's backlog, when ideas come out of your head instead of out of what buyers write.
  • You need to split a request for something new from a complaint about something broken.

When it won't work

Only a few dozen reviews a month: two requests are not demand yet.

How the agent does it

1

Set period and products

Nothing to export — the agent reads the reviews itself. Name the categories or SKUs in scope, or it takes everything and blends different products together.

2

Run the collection

The agent keeps only requests to change the product, merges different wordings of the same ask and sets aside anything mentioned once.

3

Check the list

Before you take it to the team, have the agent review its own work: which asks it inferred rather than read, and which products are too thin to judge.

What you set

Ozon and Wildberries seller accounts with review access. By hand: the period, the products or categories, and the cut-off below which an ask is a one-off.

What you'll need

Ozon
Wildberries

Starter prompt

Copy the prompt or open it straight in a chat with the agent.

Prompt for the agent

From the Ozon and Wildberries reviews for [period] collect only product change requests: what is missing, what people ask to add, what they ask to bring back. Do not take complaints about delivery, packaging and support into this list — count them on a separate line and stop there. Group the requests by meaning, not by wording. For each group: frequency, share of all reviews for the period, [3] verbatim quotes, the SKUs where it occurs. Do not turn a one-off wish into a trend: keep groups with fewer than [5] mentions in the "weak signal" section rather than in the main list. Output a table: request, frequency, share, quotes, SKUs.

Open in chat
Check the result

Check the list of requests. List: — groups with fewer than [5] mentions that you nonetheless placed higher on the list; — requests merged into one group, and the signal you merged them by; — rows where the request is actually a complaint about delivery or support; — quotes from which the request does not read directly but was inferred by you; — SKUs with too few reviews over the period to draw a conclusion from. What you can regroup yourself — show as an "I propose to change" list, leave the decision to me.

A second prompt — the agent uses it to review its own work and show what's left for you.

Related use cases