LinkedIn is the channel where the most expensive B2B relationships start, and the one where the most hours are spent on work that does not move the pipeline. Every B2B sales rep spends 4-6 hours a week on profile research, another 4-6 hours writing connection messages, another 2-3 hours drafting posts, and another 1-2 hours triaging notifications. That is 12-16 hours a week of work that is mostly mechanical — the kind of work that an agent can do without thinking, and that a human can review in minutes. The problem is that most "LinkedIn automation" tools are exactly the wrong shape: they auto-send connection requests and auto-DMs, which violates LinkedIn's User Agreement and gets accounts restricted. The right shape is an agent that drafts, classifies, and enriches — and a human who executes. In 2026, the agents are good enough to do this safely.

This guide walks through the four B2B sales workflows that ship the most value on LinkedIn, the prompts behind each one, the tools that wire them, and a workflow YAML you can run on a schedule and on demand. By the end, you have a sales workflow where the profile enrichment is instant, the connection sequences are drafted in your voice and waiting for your approval, the posts are scheduled with the best-time optimization, and the engagement digest lands in Slack before your standup.

This is one of the concrete examples of the 4-stage AI workflow pipeline. The shape is the same — trigger, plan, act, review — and the domain is the B2B social selling surface where every sales rep, every SDR manager, and every founder who is doing their own outbound already lives. It is the sibling of the customer-support pipeline and the Excel pipeline on the revenue surface, and the practical complement to the Airtable pipeline for the CRM sync that this guide is built to feed.

This article is for the B2B SDR whose pipeline is bottlenecked on research and drafting, the sales manager who wants the team to spend more time on calls and less on LinkedIn, the founder doing their own outbound who needs their week back, and the revenue ops manager who maintains the LinkedIn-to-CRM pipeline and is tired of being the integration layer.

Why LinkedIn is the right next agent rollout

Three reasons every B2B team we work with in 2026 puts LinkedIn at the top of the sales workflow automation list.

The pain is universal, visible, and paid for. Every B2B rep does 12-16 hours a week of LinkedIn work that is mostly mechanical. The work is real, the hours are billable, the alternative (a human at the keyboard for 16 hours) is the status quo. The team that cuts that to 4-6 hours without losing pipeline output has just doubled the rep's effective capacity. This is the highest-ROI sales surface an AI agent rollout can target.

The audience is the most valuable one for AI agent platforms. A B2B sales rep who reads a well-written LinkedIn automation guide is a sales rep who has budget for tools. A sales manager who finds the article on Google is a sales manager who is already evaluating. The GolemWorkers audience is the B2B sales org — founders, SDRs, sales managers, RevOps — and this article is the on-ramp for that audience.

The tool that automates LinkedIn badly is the tool that gets the team banned. This is the dark side of LinkedIn automation and the reason most tools are the wrong shape. LinkedIn's User Agreement explicitly forbids auto-sending connection requests, auto-DMs, and bulk scraping. The tools that do this get accounts restricted within weeks. The agent rollout that respects the boundary — agent drafts, human executes, every action is auditable — is the rollout that survives.

The risk is the same as any other agent rollout, multiplied: a banned LinkedIn account is a sales rep's livelihood. The way to ship this safely is the way we have been shipping every other agent in this series — start small, keep the human in the loop on every send, treat the agent as a draftsperson and never as an executor.

The 4 workflows, in order of ROI

We picked these four by combining the same three filters we use for every workflow in this series: routine (every B2B rep does them), high-leverage (the savings show up in the team's week), and tractable today (off-the-shelf tools can do them without a custom build). The order is roughly ROI, from highest to lowest, but you should still start with the one that hurts the most on your specific workflow.

The four:

1. Profile enrichment. Read every LinkedIn profile in your target list, classify seniority, function, industry, company size, intent signals, and recent activity, and push the enriched record to the CRM. The research that took 4 hours takes 4 minutes.
2. Connection-sequence drafts. For every new lead in the pipeline, draft a personalized connection request, a 1st message after acceptance, and a 3-touch follow-up sequence — all in your voice. The rep opens the drafts, reviews, and clicks send. The drafting time drops from 6 hours to 30 minutes.
3. Post scheduling. Draft a post in your voice, generate a carousel version for visual platforms, schedule it for the best-time window, and queue it for human review before publish. The rep's content cadence goes from 2 posts a week to 5 without losing the rep's voice.
4. Engagement digest. Read every mention, comment, reaction, and DM in the last 24 hours, classify into signal vs noise, and post a daily digest to Slack. The rep catches up on every interaction in 2 minutes.

Each one is one OpenClaw agent. They run as a single workflow that lives next to your CRM. Let us walk through them.

Workflow 1: Profile enrichment

The profile enrichment agent is the one that turns "I have a list of 500 prospects and I do not know which ones to call first" into "every prospect is enriched, scored, and prioritised before the rep opens the list". It is the lowest-risk entry point and the foundation for everything else on this list.

The problem: Your SDR has a target account list of 500 companies. For each, they need to find the right person to reach out to, understand the person's seniority, function, and likely pain, and write that down before the call. The research takes 4-5 minutes per prospect. For 500 prospects, that is 33-40 hours of work — a full work-week — before the first call is made. The agent does the research in 4 minutes per prospect.

Inputs:

• The target list: company names or domains, plus any seed contacts
• The enrichment source: LinkedIn Sales Navigator API, Apollo, Clearbit, ZoomInfo — pick one, or combine
• The enrichment fields: seniority (C-suite, VP, Director, Manager, IC), function (Sales, Marketing, Engineering, Product, Ops, Finance), industry, company size, recent activity (post, job change, funding event), intent signals (job posting, tech stack, hiring spree)
• The CRM target: which CRM to push to, which fields to map

Agent prompt:

You are a LinkedIn profile enrichment steward. Read every profile in the input list. For each, produce an enriched record:

>

1. Read the profile (Sales Navigator API, Apollo, Clearbit, or a manual scrape with explicit consent).

>

2. Classify the person: - Seniority: C-suite (C-level), VP, Director, Manager, IC, Unknown - Function: Sales, Marketing, Engineering, Product, Ops, Finance, HR, Legal, Other - Industry: map to the team's canonical industry list (12-20 industries max) - Company size: Solo / 1-10 / 11-50 / 51-200 / 201-1000 / 1001-5000 / 5000+

>

3. Identify intent signals: - Recent post in the last 30 days (yes / no, topic, sentiment) - Job change in the last 90 days (yes / no, old title, new title) - Funding event in the last 90 days (yes / no, round, amount) - Job posting for roles that match the team's ICP (yes / no, role, urgency) - Tech stack matches the team's product (yes / no, stack)

>

4. Compute a lead score (0-100): - Seniority match: +20 if C-suite/VP/Director - Function match: +20 if matches ICP function - Company size match: +15 if in ICP range - Industry match: +15 if in ICP industries - Intent signal: +10 per signal, max +30

>

5. Generate a 1-line summary for the rep: " is