←Attio plays
Public data

Add LinkedIn post engagers to Attio, new people only

A Kestrel build, not client work. We ran it once on one of our own LinkedIn posts, writing into our own Attio test workspace. It pulls the comments and reactions, cuts the obvious noise for free, has Jev qualify each person, resolves reactor URLs to the public profile, and checks Attio before writing. New people are created, and people already in Attio only get a list entry that says so. Nothing was sent to anyone.

Works ifYour buyers react to and comment on LinkedIn posts, and Attio is where your contacts live.

Built for · Kestrel · Attio test workspace
HarvestAPIDeeplineJevAttioTypeScript

Built asDeepline PlayPer postcapped at $0.60 a run

Get the prompt for your coding agent

The system

~/builds/attio-post-engagers
01Source

Comments

One page of comments

Replies included.

HarvestAPIDeepline
45

Reactions

One page of reactions

Member-id URLs.

HarvestAPIDeepline
67
One record per person, noise cut for free93 unique, 5 company pages cut
02Qualify

Jev

Fit, competitor or not fit

Fit counts at confidence 0.6 or more. Cached per person.

fitcompetitornot fit
Jev
4 qualified
Gate: new to Attio?public URL, member-id URL, exact nameAttio
03Write back

New

Create the person

In Attio before = false.

Attio

Existing

List entry only

In Attio before = true.

Attio
A list of qualified engagers in Attio, with the ones you already knew marked.

The signal

Someone reacted to or commented on a LinkedIn post in your space.

Why it predicts a purchase

Engagement is recent, public and on topic. The Attio check keeps people you already talk to, or already have in a deal, out of a cold sequence.

How it works

01

Pull the post's engagers

The latest original post on the profile, then one page of comments and one of reactions through Deepline's HarvestAPI tools. One record per person, and someone who commented and reacted counts once, as a comment. Company pages, open-to-work profiles, students and recruiters are cut before any model runs.

02

Qualify with Jev

Each headline and comment is sorted into fit, competitor or not fit. A person qualifies only on fit with confidence 0.6 or higher, and every answer is cached per person so a re-run gives the same list.

03

Gate: new to Attio?

Qualified reactors get their member-id URL resolved to the public profile URL. Attio People is then checked on that URL, on the member-id URL, and on an exact name match. A match counts as already in Attio only when the record existed before the run started.

04

Write back to Attio

New people are created with name, headline and LinkedIn URL. Every qualified person gets one entry on the post engagers list with the post URL, comment or reaction, and In Attio before. Someone already on the list for this post is skipped.

Build notes

  • The post was our own, and it was about outbound tooling, so 50 of the 88 engagers were sellers of outbound services. On a post aimed at your buyers the competitor share will be lower.
  • Before the Jev answers were cached, two runs on the same engagers gave 6 and then 7 fit verdicts. Caching per person plus the 0.6 confidence floor keeps borderline people from flipping in and out of the list.
  • The first version capped the list at 50 with commenters first. Commenters filled it, so none of the reactors were qualified. The cap is now 100, and Jev is cheap enough that reactors aren't cut before qualifying.
  • The already-in-Attio path was tested by deleting one person's list entry and keeping the person. The next run created no new person and wrote one entry with In Attio before set to true. A re-run after that wrote nothing.

Questions

How is this different from the post engagers skill?

The skill loads approved people into a HeyReach list. This build writes to Attio: it checks every qualified person against your People records first, adds only the new ones, and marks on the list entry whether the person was already in Attio before the post.

Why resolve the LinkedIn URL before the dedupe?

Reactors come back with a member-id URL (/in/ACoAA...), while a CRM stores the public one (/in/jane-doe). The two never match, so a person already in Attio would be created a second time. The lookup costs $0.003 and only runs for qualified people.

More builds

All builds →
Public data

Detect 163 B2B tools from a company's website and DNS

Your product replaces, integrates with or sells to users of specific tools.

Open→
Public data

Reach companies before their public deadline

Your buyers make public promises with a deadline.

Open→
Public data

Reach companies hiring for what you sell

You sell a service that companies also try to hire for.

Open→

The next step

Let's get started.