# Finding niches, Method 3: AI tools and NLP categories (/docs/niche-research/finding-niches/finding-niches-ai-tools)

---
type: Document
title: "Finding niches, Method 3: AI tools and NLP categories"
description: Using ChatGPT prompts and Google's NLP content categories to brainstorm low-competition niche and sub-niche ideas.
status: stable
order: 6
added: 2026-09-14T15:05:41Z
generated: { by: human:ahmedraza, at: 2026-11-06T18:14:10Z }
ksor:
  audience: [public]
  owner: human:ahmedraza
  approval: { by: human:ahmedraza, at: 2026-11-06T18:14:10Z }
verified: [{ by: human:ahmedraza, at: 2026-09-22T12:00:00Z }]
trust_tier: human-reviewed
build_id: sha256:401a1d4b5eead93b53bbfcc53b1ee4048e592b4be6f6059d848ed6ffaa614755
dirty: true
ksor_version: 0.0.60
---

*Taught by Sir Tanveer Nandla.*

This method skips tools entirely and uses reasoning instead: an AI model to
brainstorm candidate niches, and Google's own content taxonomy to check where
the gaps in those categories sit. Both are free, and both are best used to
generate a shortlist to verify in Ahrefs or SEMrush afterward, not as a final
answer on their own.

## 1. Use ChatGPT / AI tools

Ask targeted questions to brainstorm ideas and surface lesser-known
sub-niches. The AI needs to be told who you are and what you're doing before
it can help. A bare "find me a niche" request does not work: a model with no
role, no goal, and no constraint just returns generic, saturated suggestions,
the same ones anyone else asking the same bare question would get.

Example prompt:

> "Consider you're an SEO & Niche Blogging Expert. The goal is to find low
> competition, high search volume niches. Exclude saturated and high
> competition niches and help me find low competition niches. The sample
> niche that I want you to find is: AI background remover. You can find
> combinations of niches for me like AI [Keyword] remover. Remember, niches
> should be low competition. Low competition means low DA sites on the 1st
> SERP. I understand you can't check DA directly, but somehow, if you can
> give an idea. And only list those niches, without description."

Or, working from keywords you already have:

> "I have few combinations of niches, I am using Keyword Explorer of Ahrefs,
> so can you suggest keywords same like menu prices, [paste your keywords
> here]"

Then keep asking follow-up questions (for low-competition niches, for a
reference website, or anything else narrowing your criteria) and analyze
whatever keywords or content the AI generates rather than using it
directly.

### Other combinations to try

- **AI [Keyword] Finder**: "AI product finder", "AI job finder"
- **AI [Keyword] Generator**: "AI content generator", "AI quiz generator"

The aim is specific subtopics that big players have not heavily targeted.

A separate, reliable category of tool niche: ask ChatGPT directly for
**calculator**, **generator**, and **convertor** tool ideas: these three
tool types are a dependable source of low-competition keyword niches on
their own, independent of the AI-branded combinations above.
For a tool niche specifically, Tier 1 countries only need 15,000+ monthly
search volume to be worth pursuing, the same threshold as an
informational/blog niche (see [Niche research: tools, metrics, and
selection criteria](../overview.md)).

## 2. Use NLP categories

Where the AI-prompting method above generates ideas by reasoning, this one
generates them by browsing a fixed structure: Google Cloud's own Natural
Language content taxonomy is the same hierarchy Google's classifier uses to
file content into topics, so walking it directly shows which subcategories
exist and, by extension, which of them nobody has built a dedicated site
around yet.

Steps:

1. Open the Google Cloud Natural Language Categories page.
2. Review the hierarchical categories: most have several layers of
   subcategory nested under a broad top-level topic.
3. Match your topic to the closest relevant category, and read the
   subcategories sitting beside it.
4. Use those category keywords in your content and SEO strategy.

This helps search engines (and Google's own NLP) accurately identify and rank
your content, which improves visibility. Content that maps cleanly onto a
category Google already recognizes has an easier time being correctly
classified than content that reads as belonging to no clear topic at all.

## Worked low-competition examples

1. **Study grants & scholarships in Australia**: many people search
   scholarships broadly, but narrowing to one country narrows the field.
   Sub-niches: postgraduate scholarships in Australia, women-specific grants
   in Australia.
2. **Simulation games**: gaming overall is saturated, but sub-genres like
   farming sims or city-building sims surface low-competition keywords.
   Sub-topics: mobile simulation game reviews, PC simulation game tips.