Growth Lead
LinkedIn Influencer Discovery β Finding 100 AI Decision-Makers in 10 Minutes
Key Takeaway
An AI agent scrapes and ranks LinkedIn influencers by topic, role, and engagement β turning 4 hours of manual research into 10 minutes of automated prospecting.
The Problem
PyratzLabs needed visibility in the European AI scene. Not followers β connections. The people who decide which tools their teams adopt, which startups they invest in, which products they recommend.
Finding those people on LinkedIn manually is brutal. You search "CTO AI" and get 50,000 results. You click through profiles one by one. Is this person actually active? Do they post about AI or just have it in their title? Are they in Paris or Palo Alto? Do they have 200 followers or 20,000?
Four hours of scrolling gives you maybe 20 qualified profiles. And you haven't even started the outreach. You've just built the list.
For our visibility campaign β positioning Mr.Chief as the infrastructure layer for AI agent teams β we needed 100+ qualified targets. Manually, that's a full work week. We had a Tuesday.
The Solution
Mr.Chief's LinkedIn Influencer Discovery skill uses Apify scrapers to search, filter, and rank LinkedIn profiles by topic relevance, engagement quality, follower count, and geography. The agent takes a targeting brief and returns a prioritized outreach list with personalized angles β 100 profiles in about 10 minutes.
The Process
Step 1: Define the targeting brief
View details
Find top LinkedIn voices in:
- Industry: AI/ML, developer tools, AI infrastructure
- Roles: CTO, VP Engineering, Head of AI, AI Lead, Founding Engineer
- Company size: >100 employees
- Location: Paris, London, Berlin, Amsterdam
- Content focus: posts about AI agents, LLMs, developer productivity
- Minimum followers: 1,000
- Must have posted in the last 30 days (active creators only)
Step 2: Agent runs the discovery
bashShow code
# Agent uses linkedin-influencer-discovery skill
# Searches by topic keywords across multiple verticals
# Filters by geography, role, company size
# Ranks by composite score: followers Γ engagement rate Γ relevance
The skill queries LinkedIn's public data through Apify actors, pulling profile data, recent posts, engagement metrics, and follower counts. No LinkedIn account needed for the search β it uses public data endpoints.
Step 3: Scoring and ranking
markdownShow code
# Agent scoring model:
Score = (follower_count Γ 0.2) + (engagement_rate Γ 0.35) +
(content_relevance Γ 0.3) + (connection_overlap Γ 0.15)
# engagement_rate = avg(likes + comments) / followers per post
# content_relevance = keyword match in last 10 posts
# connection_overlap = mutual connections with our network
Step 4: Output β prioritized list with angles
markdownShow code
## Top 100 AI Decision-Makers β Paris/London/Berlin
| # | Name | Role | Company | Followers | Eng Rate | Score | Angle |
|---|------|------|---------|-----------|----------|-------|-------|
| 1 | Marie L. | VP Eng | DataCorp | 14.2K | 4.7% | 92 | Posts about agent orchestration β share Mr.Chief architecture |
| 2 | Thomas K. | CTO | AIScale | 8.9K | 6.1% | 89 | Recent post about LLM costs β our memory compression saves 40% |
| 3 | Sara M. | Head of AI | FinTech AG | 22.1K | 3.2% | 87 | Spoke at AI Summit β invite to Mr.Chief meetup |
...
Step 5: Personalized outreach prep
For each top-25 target, the agent generates a personalized connection angle:
markdownShow code
### Marie L. β VP Eng, DataCorp
- Recent posts: agent orchestration challenges, multi-model routing
- Shared connection: 3 mutual (Pierre, Yann, StΓ©phane)
- Angle: "Your post about agent orchestration resonated β we're running
31 agents on a framework I built. Would love to swap notes."
- Best approach: LinkedIn comment on recent post β connection request β DM
The Results
| Metric | Manual Research | Agent Discovery | Change |
|---|---|---|---|
| Profiles found | 20 | 127 | +535% |
| Time spent | 4 hours | 10 minutes | -96% |
| Qualified (met all criteria) | ~15 | 103 | +587% |
| With personalized angles | 0 (no time left) | 25 (top tier) | β |
| Connection acceptance rate | 18% (generic) | 41% (personalized) | +128% |
| Cost | $0 + 4h labor | ~$2 Apify credits | Negligible |
The connection acceptance rate is the real number. Generic "I'd like to connect" messages get ignored. "Your post about X resonated, here's why" gets accepted. The agent doesn't just find people β it finds the angle.
Try It Yourself
- Install the
linkedin-influencer-discoveryskill in Mr.Chief - Configure your Apify API token for LinkedIn scraping actors
- Define your targeting brief: industry, roles, geography, activity thresholds
- Run the discovery β review the ranked output
- Use the personalized angles for your top targets and track acceptance rates
Don't spray and pray. The whole point is precision. 100 targeted profiles with personalized angles will outperform 1,000 generic connection requests every time. Let the agent do the research. You do the relationship building.
I stopped guessing who to connect with. The agent tells me who's worth my time β and why.
Related case studies
Event Organizer
Event Speaker Outreach β From Luma List to Confirmed Speakers in One Week
AI agent researches target speakers, sends personalized multi-channel outreach, tracks the pipeline in Google Sheets β 3 confirmed speakers for Barcelona meetup in one week, 70% response rate.
Founder
Slack-to-Telegram Bridge β Team Messages Without Opening Another App
An AI agent bridges Slack and Telegram bidirectionally β forwarding only messages that matter to Telegram, posting replies back in Slack β zero Slack checks, 4-min response time.
Founder
Searching 3 Years of WhatsApp History β Find Any Conversation in Seconds
An AI agent indexes 3 years of WhatsApp messages and returns exact conversations in 5 seconds from natural language queries β no more 40-minute scrolling sessions.
Want results like these?
Start free with your own AI team. No credit card required.