— Quick Read

Turn a Company Name Into a LinkedIn Activity Report

This workflow shows how to build an AI agent that researches a company’s recent LinkedIn activity and turns it into a short business report. A sales, marketing, partnerships, or strategy team can use it to understand what a company is talking about, which themes show up repeatedly, and which recent posts are worth referencing.

Want the full recipe? Here is the reusable guide an AI builder can follow.

You can also try the finished agent for yourself here.

What You Are Building

You are building a company research helper. The user gives it a company name, and the agent finds the company’s LinkedIn page, collects recent public posts, and creates a markdown report with activity signals, repeated messaging themes, notable posts, and follow-up angles.

The business value is repeatability. Instead of asking one person to search LinkedIn, copy links, scan posts, and write notes by hand, the workflow turns that process into a reusable agent that can be called again and again with different company names.

What the Tools Do

  • Blocks is a network for running AI agents and connecting one agent to another. In this workflow, Blocks lets the LinkedIn company report agent be registered, called, and reused by other workflows.
  • Railway is where the agent runs continuously online. Think of it as the always-on home for the agent.
  • Apify is a service that collects structured information from websites and platforms. Here it collects public LinkedIn company posts and returns them as a dataset.
  • DuckDuckGo, SerpAPI, or Bing Search help the agent find the right LinkedIn company page from a company name.
  • GitHub is the place where reusable instructions and project documentation can be stored.
  • An API key or token is a private password-like credential that lets the agent use a service such as Blocks, Apify, SerpAPI, or Bing.
  • A CLI, or command-line interface, is a builder tool used to validate, deploy, register, and test the agent.
  • A dataset is the structured result returned by a research run. In this workflow, it contains the LinkedIn posts that the report is based on.

What the Guide Gives You

The linked guide gives an AI builder enough detail to reproduce the workflow. It covers the agent inputs, environment variables, Blocks card, Railway deployment, Apify actor setup, validation steps, and safety notes.

It is meant for a technical agent or builder. This page is the business-readable version.

Before You Start

  • Blocks account, so the agent can be registered and called by other agents or workflows.
  • Blocks API key, so the provider can authenticate while it runs.
  • Railway account and project access, so the agent can stay online as a worker.
  • Apify account and API token, so the agent can collect recent public LinkedIn posts.
  • Access to the Apify LinkedIn actor harvestapi/linkedin-profile-posts, because that actor powers the LinkedIn post collection.
  • Optional SerpAPI or Bing Search key, if you want a controlled search provider instead of the built-in DuckDuckGo fallback.
  • GitHub access, if you want to publish or update the reusable skill file.
  • Node.js and npm, if you are rebuilding or modifying the project locally.

No OpenAI API key is required for this specific version. The report is created from the collected LinkedIn evidence by deterministic project code.

How It Comes Together

First, the user submits a company name such as “OpenAI.” The agent turns that into a search query and looks for a LinkedIn company page.

Next, the agent validates that the selected link is a LinkedIn /company/ page. If the user already knows the correct LinkedIn URL, they can provide it directly and skip the search step.

Then Apify collects recent public posts from that LinkedIn page. The workflow keeps comments and reactions off by default to control cost and keep the output focused.

Finally, the agent creates two outputs: a markdown report that a business user can read, and a JSON evidence file that another system or builder can inspect later.

Starter Prompt

Use this reusable skill file to build or adapt a LinkedIn company report agent:
https://raw.githubusercontent.com/CagedEther/agentic-business-schooled/refs/heads/main/linkedin-company-report-agent-skill.md

Build the agent for this use case:
- Company report input: [COMPANY NAME OR CRM FIELD]
- Default post window: [6months, month, year, or any]
- Max posts: [NUMBER]
- Hosting target: [Railway or another worker host]
- Blocks agent name: [AGENT_NAME]
- Required output format: markdown report plus JSON evidence

Keep secrets out of source control. List all required accounts, API keys, and deployment variables before implementation.

What a Good Result Looks Like

A good result gives a clear company snapshot, the LinkedIn page used, how many posts were analyzed, repeated themes, notable posts with links, and practical follow-up angles. The report should be concise enough for a sales or marketing teammate to scan quickly.

The JSON evidence should make the report auditable. It should show the search provider, selected LinkedIn URL, Apify actor, dataset ID, normalized posts, and any warnings.

Use It Responsibly

This workflow depends on public web data. LinkedIn pages can change, posts can be removed, and scraping tools can return incomplete data. Treat the output as research support, not as a guaranteed source of truth.

Watch costs when increasing post limits or enabling comments and reactions. Those options can make Apify runs larger and more expensive.

Do not put private credentials, customer records, internal strategy notes, or confidential account data into public documentation or logs.

What to Try Next

Connect the company report agent to a CRM or account-planning workflow. For example, when a salesperson opens a target account, another agent could call this report agent, summarize the latest LinkedIn themes, and draft three account-specific outreach angles for human review.