Platform: Instagram
├── "{niche} influencers Instagram 2024 2025 followers"
├── "{niche} content creators Instagram {region}"
└── "top {niche} Instagram accounts followers"

Platform: TikTok
├── "TikTok {niche} creators followers 2024"
├── "{niche} TikTok influencers {region}"
└── "best {niche} TikTok accounts"

Platform: YouTube
├── "YouTube {niche} channels subscribers 2024"
├── "{niche} YouTubers {region}"
└── "{niche} YouTube creators"

Platform: X/Twitter
├── "Twitter {niche} advocates influencers followers"
├── "{niche} experts X Twitter"
└── "{niche} thought leaders Twitter"

Platform: Facebook
├── "Facebook {niche} pages groups followers"
├── "{niche} community Facebook {region}"
└── "{niche} organizations Facebook"

Key: Run ALL searches in parallel (single message with multiple WebSearch calls).

Step 2: Deep Fetch Aggregator Sites

Fetch structured data from curated lists:

Site	URL Pattern	Data Quality
Feedspot	`influencers.feedspot.com/{niche}_instagram_influencers/`	High - top 100 lists
The Influence Agency	`theinfluenceagency.com/blog/{niche}-influencers-to-follow`	High - curated with handles
Upfluence	`upfluence.com/find-influencers/top-{niche}-influencers`	Medium - regional focus
IZEA	`izea.com/resources/content-creators-with-{niche}/`	Medium - brand partnership focus
Heepsy	`heepsy.com/top-{platform}/{niche}`	High - platform-specific

Step 3: Data Normalization

Cross-reference sources to:

Verify follower counts (average if sources differ >20%)
Correct handles (many sources have typos)
Fill gaps (combine name from one source, handle from another)
Deduplicate (same person across multiple sources)

Step 4: Excel Generation

Generate formatted Excel with Python:

import pandas as pd
from openpyxl.styles import Font, Alignment, PatternFill

# Standard columns for all platforms
columns = [
    'handle', 'name', 'followers', 'country', 'platform',
    'niche', 'description', 'profile_url', 'engagement', 'content_type'
]

# Create separate DataFrames per platform
# Combine into single Excel with:
# - One sheet per platform
# - "All_Platforms" combined sheet
# - Formatted headers (blue fill, white text)
# - Auto-sized columns
# - Wrapped text for description

output_path = f'{baseDir}/output/{Region}_{Niche}_KOLs.xlsx'

Step 5: Output Summary

Provide chat summary with:

File path
Count per platform
Top 3 influencers per platform (by followers)
Data completeness %

Data Schema

Field	Description	Example
handle	Platform username with @	@thepsychdoctormd
name	Display name	Dr. Sasha Hamdani
followers	Count with K/M suffix	936K
country	ISO country or region	USA, Canada, Singapore
platform	Source platform	Instagram
niche	Content focus	ADHD Psychiatrist
description	1-2 sentence bio	Board-certified psychiatrist, author...
profile_url	Direct link	https://instagram.com/thepsychdoctormd
engagement	High/Medium/Low	High
content_type	Content themes	Medical education, Q&A

Efficiency Principles

Principle	Implementation
Parallel execution	Multiple WebSearch in single message
Leverage aggregators	Use curated lists, not individual profile searches
Single script output	One Python block creates entire formatted Excel
No manual steps	Fully automated end-to-end

Example Invocations

Example 1: User says "find me travel influencers in Southeast Asia"

Niche: travel
Region: Southeast Asia
Platforms: All
Target: 50/platform
Output: SEA_Travel_KOLs.xlsx

Example 2: User says "I need ADHD creators for a partnership campaign"

Niche: ADHD
Region: North America (inferred from English request)
Platforms: All
Target: 30-50/platform
Output: NorthAmerica_ADHD_KOLs.xlsx

Example 3: User says "TikTok beauty influencers in US"

Niche: beauty
Region: USA
Platforms: TikTok only
Target: 50
Output: USA_Beauty_TikTok_KOLs.xlsx

Integration with Other Skills

Chain with:

/founder-content — Create outreach message templates
/linkedin-gtm — Plan engagement strategy for LinkedIn KOLs
/twitter-x-gtm — Plan engagement strategy for Twitter KOLs
/event-gtm — Identify KOLs attending same conferences

Autonomous Mode

When user says variations of "don't ask me" or "work autonomously":

Make reasonable assumptions for missing parameters
Default region based on language (English → North America, Chinese → China/SEA)
Default to all 5 platforms
Default to 30-50 per platform target
Proceed without confirmation prompts

Parameter	Required	Default	Examples
Niche	Yes	-	"ADHD", "travel", "DTC beauty", "AI/tech", "fitness"
Region	Yes	-	"North America", "Southeast Asia", "Europe", "Global"
Platforms	No	All 5	Instagram, TikTok, YouTube, X/Twitter, Facebook
Target Count	No	30-50/platform	Any number

influencer-research

Influencer Research Skill

Inputs

Workflow

Step 1: Parallel Web Research