Reach people engaging with competitor posts
A scheduled scraper watches the LinkedIn posts of 15 competitors, collects every reaction and comment, enriches the people, and runs every person through an AI agent that checks ICP fit.
Works ifYour buyers follow competitors or category voices on LinkedIn.
The system
Competitor A
Latest posts
Competitor B
Latest posts
+ 13 more
Latest posts
Cascade
Best provider first
ICP agent
Thinks, then searches
Picks between two search tools before deciding fit.
Reaction
Liked a competitor post
Comment
Said something about it
The signal
Someone reacted to or commented on a competitor's post.
Why it predicts a purchase
They are paying attention to the category right now, and they told you which vendor they follow. That is a topic signal and a competitive signal in one click.
How it works
One key per reactor per post
Dedup on post plus reactor, so ten people reacting to one post become ten signals and the same person on the next run does not.
Best provider first, fallbacks after
A three-step cascade: the best-quality provider first, a backup API second, raw scraping last.
An agent inside a normal workflow
The ICP check is a real agent with a think step and two search tools that it chooses between, living inside an otherwise deterministic pipeline.
Build notes
- The tracker was built and then never wired to a trigger. It sat idle for two weeks. Check that a workflow runs, not only that it exists.
- Signal types follow a fixed taxonomy: reaction, comment, profile_visit.
- Inline JSON in SQL broke five runs in a row. Parameterised queries only.
Questions
Should I mention that they liked a competitor's post?
No. Mention the topic, never the like. The engagement tells you who is paying attention, not what to say.
Are comments better than likes?
We would weight them higher: comments often state the pain in their own words.
More builds
All builds →Turn a competitor's case studies into leads
Your competitors publish case studies or logo walls.
Find out which competitor tool they run
You replace a tool your buyers already run.
Detect 163 B2B tools from a company's website and DNS
Your product replaces, integrates with or sells to users of specific tools.
The next step