Bulk Judge · Free for ChatGPT and Claude
Know what every customer said. 5,000 rows in under a minute.
Drop in reviews, search queries or ad comments. See what people say most, counted, in their own words.
Not listed yet · Leave your email for the link
Example · Ad comments, sorted by theme
AI-generated labels
| Comment | Category |
|---|---|
| $149 for a blender? Mine from Target works fine. | Price |
| Every review on their site is five stars. Feels fake. | Trust |
| Will it actually crush ice or just make noise? | Does it work |
| Love it, but not at full price. Waiting for a sale. | Price |
| Would buy if it came with a longer warranty. | TrustClose call |
| Had it six months. Smoothies every morning, zero complaints. | Praise |
| Cheaper ones exist. What makes this one worth it? | Price |
| DM me for free gift cards | Spam |
- Price3
- Trust2
- Does it work1
- Praise1
- Spam1
Why it stays unread
The answer is in the pile. Nobody has time to read the pile.
- 01
Reading it yourself takes days.
A few hundred rows in, you start skimming, and the patterns you were looking for blur into each other.
- 02
Pasting it into chat stops partway.
Give a chat assistant a big sheet and it summarizes the start. In our test on 5,000 rows, chat sorted 70% and stopped.
- 03
The objection that costs you sales is easy to miss.
It is rarely the loudest comment, so it only shows up once every row has been counted.
What you find out
Whatever is piling up, find out what it says.
See which search queries are buyers. Find the objection keeping your ads from converting. Know what customers keep praising before you write the next campaign.
| You have | You find out |
|---|---|
| Customer reviews | What people praise, what they complain about, and how often each comes up. |
| Search queries | Which searches come from buyers and which are just research, so spend follows intent. |
| Ad comments | The objection that keeps coming back, like the price or “does it actually work?” |
| Support tickets | Which tickets are urgent, before anyone has read them all. |
| Leads | Which leads look like your best customer, and which do not. |
How it works
Works inside the chat you already use.
- 01
Hand it the sheet.
Attach a file or paste the rows, up to 5,000 at a time.
- 02
Say what to sort by.
Name your own categories, or pick a ready-made set for reviews, search queries, ad comments, support tickets or leads.
- 03
Check the first rows.
You see the first 200 labels before it does the rest, so a wrong category gets fixed before it spreads to every row.
What comes back
Every category is assigned by AI. The close calls are marked so you check those first.
- How often each one comes up
- A count for every category, so the loudest complaint and the most common one stop looking like the same thing.
- A real example of each
- One row per category in the customer’s own words, ready to quote in a meeting or a brief.
- The close calls, flagged
- The rows it is least sure of are marked, so you know where to look first.
- Your sheet back, sorted
- Your spreadsheet with a category on every row, to filter, chart or hand to someone else.
The test
We tested it against pasting the sheet into chat.
We took 1,800 rows that people had already sorted by hand, then gave the same sheets to a leading AI chat assistant. Bulk Judge matched or beat it on every sheet, and sorted every row.
On a 5,000-row sheet the gap widened. Chat sorted 70% and stopped. Bulk Judge finished all 5,000 in under a minute.
| Sheet | Chat alone | Bulk Judge |
|---|---|---|
| Support tickets600 rows | 99.5%sorted 99.8% | 99.5%sorted 100% |
| Product reviews600 rows | 92.7%sorted 100% | 92.8%sorted 100% |
| Search queries600 rows | 48.5%sorted 99.7% | 48.5%sorted 100% |
| News headlines5,000 rows | 63.4%sorted 70.3% | 86.6%sorted 100% |
Each figure is the share of rows given the same category a person gave, comparing chat’s best run with Bulk Judge’s weakest.
Search intent is hard for everyone: both agreed with people about half the time.
On the rows it finished in the 5,000-row test, chat was more accurate per row (90.2% against 86.6%). It ran out before the end.
Done for you
Want us to turn it into a report?
Send us the sheet. We sort it, read it, and hand back what your customers are actually saying and what to do about it, as a customer insight report or a search intent audit.
Questions
Fair questions.
Is it really free?
Yes. Bulk Judge will be free to use inside ChatGPT and Claude. If you would rather have a person read the results and write up what they mean for your business, that is a separate paid service from Mather Media Solutions.
How accurate is it?
In our test against 1,800 rows that people had already sorted, it matched or beat a leading AI chat assistant reading the same sheets, and it sorted every row. It is not perfect: some categories are genuinely hard, and on search intent both tools agreed with people only about half the time. The categories are assigned by AI, so check the close calls it flags.
What happens to my data?
Your rows are sent to an outside AI service to be sorted, and the results are deleted after 24 hours. Do not upload payment card numbers, health records, government IDs or passwords.
How big a sheet can it handle?
Up to 5,000 rows at a time. It sorts the first 200 first so you can check the categories make sense, then does the rest. Once you approve the categories, a 5,000-row sheet takes under a minute.
Can I choose my own categories?
Yes. Name your own, or start from ready-made sets for reviews, search queries, ad comments, support tickets and leads.
When can I start?
It is not out in ChatGPT or Claude yet. Leave your email on this page and we will send you the link the day it launches.
Early access
It is free. Get it first.
We send the link to your inbox the day it goes live in ChatGPT and Claude, so you can run your own sheet first.