Competitor's Refugees

The source monitors three places at once: your competitor's own LinkedIn company posts, a keyword set that catches the grumbling elsewhere — "{tool} pricing", "{tool} credits", "{tool} alternative" — and the low-star reviews on the public platforms. It runs continuously, because the signal isn't a fixed database that empties as you work it.

It regenerates. Every time your competitor ships something unpopular, the comment section refills — which means the well is topped up by the one party you don't control and can always count on: the competitor themselves.

An AI pass reads each complaint and outputs three things: the person, their company, and the pain stated in their own words. Then it files the pain into a category — pricing, billing, data quality, support silence, a missing feature — so the campaign can answer like with like instead of blasting one message at every kind of unhappy.

That categorisation is what makes the next steps possible. A billing surprise and a missing feature are two different conversations, and the only way to have the right one is to know which you're looking at before a word is written.

In every other recipe in this catalogue the person is the last and most expensive step — a LinkedIn search, an AI judge, an email finder, a bounce risk at the end of it all. Here that whole stage disappears, because the commenter already is the LinkedIn profile. The contact is not something you resolve after the signal; the contact is the signal.

That inversion is the quiet advantage of the source. The identity arrives verified and native, so the budget and the risk that usually pile up on the last step simply aren't spent.

The category from the extract step decides the split. Profile A named a grievance — the opening answers it directly, the day after it went live. Profile B showed intent without anger — asking for an alternative, missing a feature — and the opening anchors on the use case closest to their business rather than a wound they never described.

Send the wrong one and it lands wrong. Open with sympathy for a problem a Profile B prospect never had and you've invented a grievance for them; skip the answer a Profile A prospect is openly demanding and you're just another pitch in the pile.

An explicit gate before a single message is composed. Out go the non-buyers the extract step didn't already catch, the companies that fall outside your ICP however loudly they complained, the rival vendors trawling the same comment section, and the complaints that turn out to have already been resolved.

As in every recipe, the discards are read. If a keyword monitor keeps importing the same off-target crowd, that's a term to retune at the watch step — not a rejection to repeat on every sweep.

A second AI pass writes the message on a strict priority: answer their exact pain if they voiced one; if not, anchor on the use case closest to their business; and only failing both, a soft opener. The message meets the prospect where their frustration actually is, at the level of specificity they gave you.

The complaint itself is never repeated back. The message just happens to solve the thing they're angry about — it doesn't remind them they ranted about it in public, or make them feel watched.

Delivery is native LinkedIn steps, paced under invite quotas so the outreach never outruns what an account can safely send, with replies handled in the same workspace the sourcing ran in. The person who complained on Monday hears from you on Tuesday — no export, no hand-off, no cooling-off period that lets the moment pass.

Timing is half the conversion. The complaint is freshest the day it's posted; a week later the anger has settled or a competitor already answered. Quota-pacing keeps the account healthy without letting the window close.