// LINKEDIN LEAD GENERATION
Signal-Based Prospecting: Turn LinkedIn Likes Into Leads
Most outbound starts from a static list: title, company size, region. Signal-based prospecting starts from what people do: they like a post, comment on it, change jobs, announce they are hiring. A signal tells you who cares about the problem right now. Fit still matters, so you score every signal against your ICP.
Reactions versus comments
| Signal | Strength | What it gives you |
|---|---|---|
| Reaction | medium | a wide pool of people interested in the topic |
| Comment | high | their own words, the best possible opening line |
Pull both and score commenters slightly higher.
How to find engagers at scale
Doing it by hand means opening each post and scrolling through every reaction. With AtomSell your agent searches posts by topic, pulls every reaction and comment into a table, and keeps one row per person even when they engaged with several posts.
Score first, so you do not spam
- Free rules remove obvious non-fits: students, recruiters, job seekers, vendors.
- AI scores the rest 1 to 10 against your ICP, with a reason.
- Only strong fits get a note.
Turn the signal into a message
Reference the exact post, ideally the person's own comment. "Liked your point under Sylvain's post that system thinking now matters more than Python in GTM" works. "Hi, I help companies like yours" wastes the signal.
Common pitfalls
- Sending beyond safe daily limits. AtomSell paces every action for you.
- Generic notes that ignore the post.
- Contacting the same person twice from different posts.
- No success metric defined before you start.
The result
Our post-engager campaign: 2,499 requests, 81% accepted, 46.5% of accepted connections replied. Start with three to five posts, measure, then put it on a daily schedule.
On your current stack
Most stacks cannot find engagers by topic at all. Apollo and Sales Navigator lists carry no engagement data. Clay's post audiences only cover posts about your own brand, which helps PR, not prospecting. Scraper actors can pull posts, reactions and comments, but each is a separate tool on your session cookie, built by a different developer, and they break when LinkedIn changes. Teams that try it spend more time fixing the pipeline than talking to leads, so most never track live signals at all and keep sending to cold lists.
With AtomSell signals are just another search: posts by topic, keyword, author or date, everyone who engaged, scored and messaged from one table.
Run it yourself
Start a 3-day free trial with full access at app.atomsell.ai/register, connect AtomSell to Claude, ChatGPT or Cursor, and run the playbook on your own buyers.