Grok BotSalesFinding leadsAI score · 76/100AI-assisted edit · no obvious prompt-injection riskX

Triage LinkedIn DMs with Grok Bot

Input: unread LinkedIn DMs → Processing: Grok Bot reads threads and profiles, scores against ICP → Output: Slack alerts with copy-ready reply drafts

Some experienceStepsTip
Who
Sales or marketing professionals handling a high volume of LinkedIn DMs
You get
Author report: scored 128 DMs this week, drafted Slack replies for 10 worth answering, logged and skipped 118; a human sends every reply
You need
Slack

Constraints

Configure a read-only workflow in Grok Bot: read LinkedIn unread DMs newest first, extract the full thread and sender profile, score against an ideal customer profile, post A/B leads to Slack with reply drafts; never send, react, or archive on LinkedIn. Requires LinkedIn account access and a Slack workspace. Cost conditions unknown.

What they did

The author built a read-only LinkedIn DM triage workflow using Grok Bot. Grok Bot reads unread DMs newest first, extracts the full conversation thread and sender profile, scores against an ideal customer profile, posts A/B leads to Slack with operator-style reply drafts, and never touches send, react, or archive on LinkedIn.

How it went

Author report (author_report): processed 128 LinkedIn DMs this week, 10 scored as A/B and generated Slack alerts with reply drafts, 118 were logged and skipped. Every actual reply is sent by a human.

Worth knowing

When company size is unknown, those points stay at 0 and require human review. If the run hits a LinkedIn login or captcha screen, it stops. The author recommends starting with your last 20 unread DMs and writing down why each is or isn't a fit to calibrate your first rubric. No copyable prompt was provided by the source; configure the steps according to Grok Bot's agent interaction model.

Steps

  1. Read unread DMs

    Start a task in Grok Bot to read LinkedIn unread DMs newest first. For each DM: fetch the full thread (including the lead magnet DM they are responding to), extract the sender's profile (title, company, size, industry). After reading, mark the thread unread again so nothing appears touched.

  2. Score against ideal customer profile

    Score each sender against preset criteria. In: founders, CEOs, CMOs, VPs of Growth, Heads of Marketing at mid-market companies, often 50–500 employees. Out: vendors pitching the user, job seekers, recruiters, solo creators, tiny SMBs. If company size is unknown, those points stay at 0 and a person checks. A is 70+, B is 50–69, everything else gets logged and skipped.

  3. Post A/B leads to Slack

    Send A and B leads to Slack. Each alert is max 4 lines: who they are, what they said, one fit concern, links. Place a copy-ready reply draft under the thread.

  4. Draft replies like an operator

    Grok Bot drafts reply drafts for each A/B lead with these rules: if they ask about the service, answer in the first line; if they say thanks, ask one question about whether they have used it yet; if pricing comes up, reply 'it depends on scope' plus a call, never a number; one question per reply, under about 400 characters.

  5. Stay read-only, never auto-act

    Grok Bot never sends, reacts, or archives anything on LinkedIn. If a login or captcha screen appears during the run, the task stops and waits for human handling. All final replies are sent by a human on LinkedIn.

Prompt you can copy

Reference prompt · Editorial suggestion。Confirm the destination, input variables and output format before use.

This block is the prompt. Ads are never placed here.

No prompt needed — just follow the steps.

Test by having Grok Bot read your last 20 unread DMs and write down why each is or isn't a fit. That's your first rubric.

Source

Grok Bot 分拣 LinkedIn 私信：读全线程与发信人资料、按理想客户画像打分，只为 A/B 线索在 Slack 起草回复（只读不发） · ericosiu · X · 2026-10-25T21:00:07.000Z

Comments & feedback

Tried this method? Share results, problems or improvements. Comments appear after moderation. Please avoid sensitive information.

Loading…

    Related