AI Class Search
Describe a product or service in plain English and instantly get the correct Nice Classification term and class — including whether the item is a good or a service. It bridges the gap between how clients describe what they do and the exact wording the trademark registry accepts.
What is the AI Class Search?
Clients describe their goods and services in their own words — "leather wallets", "mobile phone covers", "cotton t-shirts". Those words rarely match the official accepted terminology in the Nice Classification list word-for-word, so finding the right pre-approved term for each item (and the right class) is slow, manual work.
The AI Class Search does it for you. It reads the description, understands the meaning of each item, and returns the closest officially accepted Nice terms — grouped by class — so you can copy filing-ready wording in seconds.
Plain English in, the right class out
The real value shows up when products are described in everyday, non-legal language. For each item the AI works out three things at once:
- The correct class — which of the 45 Nice classes the item belongs in.
- The correct Nice terminology — the exact accepted wording the registry expects, not the client's phrasing.
- Whether it's a good or a service — the distinction that trips up the most filings.
| "mobile phone covers" | → Class 9 — goods (covers for smartphones) |
| "online trademark search software" | → Class 42 — software service |
| "cloud-based docketing system" | → Class 42 — service |
| "retail store selling cosmetics" | → Class 35 — a retail service, not Class 3 goods |
| "nail polish" | → Class 3 — goods (accepted term nail varnish) |
Why it finds terms a keyword search misses
A plain keyword search only matches the exact words you type, so it misses accepted terms that mean the same thing in different words. Because the AI Class Search works on meaning, it surfaces the right accepted term even when the wording is completely different — "nail polish" still finds the accepted term "nail varnish", and "mobile phone covers" finds "covers for smartphones" despite sharing no words at all.
How to Use It
-
1
Paste the client's description
Drop in the goods/services exactly as the client wrote them. Separate items with commas, semicolons or new lines. Up to 25 items per run.
-
2
Optionally limit to specific classes
If you already know the relevant classes, enter them (e.g. "18, 25") to focus the results. Leave blank to search all 45 classes.
-
3
Click Match terms
Results appear grouped by class. Each accepted term shows a match score and a confidence label, along with the client phrase it was matched from.
-
4
Review cross-class flags
A notes panel highlights goods that can legitimately fall in more than one class (e.g. belts — Class 18 if leather, Class 25 if clothing) so you can confirm the client's intent.
-
5
Copy the terms you want
Use the copy button on any term, or "Copy all" for a whole class, and paste the accepted wording straight into your filing.
Why it only returns real accepted terms
Every suggested term comes from the official Nice Classification list that also powers Dexin's Class Search. The AI's job is only to select and group the best-matching accepted terms — it cannot invent wording that isn't on the official list. That means what you copy is always genuine accepted terminology, not an AI guess.
Credits & Access
| Cost | 10 credits per run. Charged once per run, and only when matches are produced. |
| Items per run | Up to 25 goods/services items. |
| Class filter | Optional. One or more classes from 1–45. |
| Access | Logged-in users. Find it under AI Tools → AI Class Search. |
Try the AI Class Search
Turn a client's plain-language list into accepted Nice terms in seconds.