LinkedIn

Automate remembering, not sending

Clay writes the note. Slack queues one card. You click Connect. Twenty minutes a day. Nothing bots LinkedIn.

The loop

  1. 01Import the list you already paid forSprouts into Clay. Do not re-enrich. Name, LinkedIn URL, department, title, company. He ran this on a list he already had.
  2. 02Write a note under 200 charactersClay formula, zero credits. No pitch. If department is blank, fall back to company so you never get an empty ‘work in ___’ line.
  3. 03Draft a one-question first messageClay AI column: no pitch, no product, no links, reference their role, exactly one easy question. 0.2 credits a row. Test five before the batch.
  4. 04Sheets is the state machine. Make fires one.Status Queued/Sent, Date Sent, Accepted, Msg 1 Sent. Make searches Queued with limit 1 so cards come one at a time.
  5. 05Slack the card. A human clicks Connect.Paste-ready note, LinkedIn URL. You open the profile and send. Accepted rows get the first message the same way. Nothing hits LinkedIn via API, bot, or plugin.
linkedin.com ↗Twenty minutes a day. You still click Connect.Never bot-send on LinkedIn. Reuse a Sprouts list in Clay, write a under-200-character connection note with a formula, draft a no-pitch one-question first message, park state in Google Sheets, and let Make fire one Slack card at a time. A human clicks Connect and sends the follow-up. Twenty minutes a day. Nothing touches LinkedIn via API, bot, or plugin.

Most LinkedIn ‘automation’ is a bot that gets the account restricted. Tharun’s version automates the memory: who is queued, what to paste, who already accepted. The click stays human. Twenty minutes a day is the constraint, not a slogan.

The interesting gate is the one-at-a-time card. Make limit 1, so you never get a dump of fifty connects. Notes stay under 200 characters. The first message asks one question and does not pitch.