— Pulling LinkedIn-active company profiles with Apify

Pulling LinkedIn-active company profiles with Apify

If you are learning how AI agents can help with business research, this is a useful starter workflow: take a company LinkedIn page, pull a public list of people associated with that company, then identify which of those people appear to be active on LinkedIn.

This is not about scraping private data or collecting emails. The goal is simpler and more practical: create a public, evidence-backed list of company employees who might be good candidates for employee advocacy, sales research, recruiting research, partnership mapping, or content collaboration.

The reusable skill file for this workflow is here: LinkedIn company profiles pull skill.

What you are building

You are building a repeatable research process, not a giant app.

The workflow should answer questions like:

  • Who publicly lists this company on LinkedIn?
  • What title or role does each person show?
  • Which profiles have recent public LinkedIn posts?
  • Who looks active enough to include in a content, sales, or advocacy project?
  • What files and source links prove where the answer came from?

This kind of workflow is a good example of agentic business work. The agent is not just writing a paragraph. It is using a tool, collecting data, normalizing messy results, applying judgment, and producing a usable output.

What Apify does

Apify is a platform for running web data collection tools called actors. For this task, the useful actors are LinkedIn-focused actors that can pull public company employees and public profile posts.

The workflow usually uses two types of actors:

  • A company employees actor to pull people associated with a LinkedIn company page.
  • A profile posts actor to check whether those people have recent public posts.

The important word is public. If a profile does not expose posts publicly, the output should say that no public activity was found by the actor. It should not claim the person is inactive.

What the skill file gives you

The skill file is a guide you can give to an AI coding agent or research agent. It tells the agent how to run the workflow safely and consistently.

It includes:

  • Which Apify actors to start with.
  • What inputs to use for the company roster pull.
  • How to avoid email or phone enrichment.
  • How to normalize names, titles, locations, LinkedIn URLs, and current-company fields.
  • How to check recent public profile activity.
  • How to classify people as active, strongly active, inactive/private, or unverified.
  • What CSV and Markdown files to produce.

You can open the skill here: linkedin-company-profiles-pull-skill.md.

Before you start

You need three things:

  1. A company name.
  2. The company’s LinkedIn page URL.
  3. Access to Apify with an API token.

For example, the company page URL might look like:

https://www.linkedin.com/company/example-company/

You should also decide what “active” means before you run the workflow. A simple beginner definition is:

Active = at least one public post, repost, or quote post in the last 6 months.

That definition is not perfect, but it is clear. Clear beats clever when you are building a repeatable research workflow.

The basic workflow

Start by asking the agent to pull the public company roster from LinkedIn through Apify. The output should be saved as a raw JSON file and a cleaned CSV file.

Next, ask the agent to inspect the roster. It should keep useful fields like name, title, company, location, LinkedIn URL, and current-position evidence.

Then run a second Apify pass against each LinkedIn profile URL to look for recent public posts. Keep the post limit small at first, such as 3 to 5 posts per person. This keeps the workflow cheaper and easier to inspect.

Finally, ask the agent to create two finished outputs:

  • A CSV of active employees with their profile links, post counts, last post date, and sample post URLs.
  • A short Markdown report explaining the method, totals, caveats, and source files.

Starter prompt

Here is a prompt you can adapt:

Use this skill file:
https://raw.githubusercontent.com/CagedEther/agentic-business-schooled/refs/heads/main/linkedin-company-profiles-pull-skill.md

I want to find employees at [COMPANY NAME] who appear publicly active on LinkedIn.

Company LinkedIn URL:
[PASTE COMPANY LINKEDIN URL]

Use Apify to:
1. Pull the public company roster.
2. Normalize the roster into a CSV.
3. Check recent public posts for each profile.
4. Classify people as active, strongly_active, inactive_or_private, or unverified.
5. Create a short Markdown report and an active-employees CSV.

Constraints:
- Do not request email or phone enrichment.
- Use a 6-month activity window.
- Keep post checks to 5 posts per employee.
- Save raw data and dataset IDs for auditability.
- Do not claim someone is inactive if the actor simply returns no public posts.

Before running the full workflow, run a 5-person test and show me the output shape.

The last line matters. A small test run helps you catch actor schema changes, bad inputs, or confusing output before spending more time or money.

What a good result looks like

A useful final report should not be fancy. It should be clear.

It should tell you:

  • Which company page was used.
  • Which Apify actors were used.
  • How many people were found.
  • How many profiles were checked for activity.
  • How many people looked active in the chosen window.
  • Which profiles could not be verified.
  • Where the raw and cleaned source files are saved.

The CSV should be easy to sort and filter. For example, a marketing team might filter for executives, sales leaders, developer advocates, or product leaders. A recruiting team might filter by geography or title. A founder might look for people who already post publicly and could become early champions for a campaign.

Use it responsibly

This workflow is powerful because it turns messy public web data into a structured business asset. That also means you should be careful.

Do not enrich the file with private contact details. Do not imply that a person is inactive just because no public posts were returned. Do not treat the actor output as perfect truth. LinkedIn profiles change, visibility settings vary, and scraping tools can return incomplete data.

The right standard is evidence-backed, not omniscient.

What to try next

Once you can pull active profiles for one company, try comparing two or three companies in the same category. You could look at which company has more active executives, which roles post the most, or what topics show up most often.

That is where the business value appears. The first workflow gives you a list. The next workflow turns that list into insight.