Independent LinkedIn Manager
Audience: front-end end users
Scenario
The company uses LinkedIn to post, keep the account going, and also to recruit and talk with talent. When nobody owns the account, posting, reviewing applicants, and organizing candidates pile up.
The system already provides a standard Independent LinkedIn Manager employee (application name linkedin01). You do not need to create an employee from scratch: click Use in the AI Employee Market to copy it into the current project, write memory and create a session. After brand, posting rules, and job requirements are in memory, you do not have to repeat them every time. It uses browser tools to manage the company’s LinkedIn account, and to help with posting and recruiting. In the left AI Employee, say in natural language what to post, who to hire, or which candidates to organize.
Write memory first, log in to the company’s own LinkedIn, then tell it what to do.
Work type
Account work and recruiting on the LinkedIn channel:
- Manage the LinkedIn account, publish posts, and do other page actions
- Publish and update job posts
- Review applicants, organize public candidate profiles, and do a first-pass screen
- Search for potential candidates by role, and help with early contact and follow-up
- Hand the organized results to HR
It does channel execution. Whether to publish, interview, or hire is still decided by the company and HR.
Configured tools
This work is done by chatting in the left AI Employee. It does not use external chat hosting. You cannot chat from My employees. The LinkedIn account needs a login. If the session does not already have a logged-in account, it opens LinkedIn in a headed browser, gives you the browser address, and you log in before it continues. Log in once. Do not tell the AI the account or password. Next time it signs in by itself.
| Setup | Purpose |
|---|---|
| AI Employee → Enter market → Use | Copy the standard Independent LinkedIn Manager (linkedin01) into the current project |
| ⋯ → Enterprise Docs | First create the LinkedIn Admin folder; open that folder, then add the memory Markdown; after create, click ⋯ → View info on the file and copy the url |
| AI Employee, Agent session | Who does the work, and which posting or recruiting task is in progress; paste the memory url into Session prompt, and say it owns and updates that memory |
| Chat box paperclip | Add attachment: Audio, Image, Camera, File (job notes, brand materials, and similar) |
| Mapped browser tools | Open and operate LinkedIn (posts, jobs, and similar; follow what is mapped to this employee) |
Map extra tools, Skill, SOP, MCP, and enterprise resources when you need a closer fit to company posting or hiring rules.
Skill and SOP
The market copy of Independent LinkedIn Manager already has basic LinkedIn-operation ability. The company can map more Skill, SOP, and materials to match its own posting and hiring process.
Logic in brief
Write memory inside the LinkedIn Admin folder, copy the url → paste the url into Session prompt so it owns and updates that memory → If the session is not logged in, open a headed browser, give you the address, you log in to LinkedIn, close it, then open again to verify → Say what to post, which job to publish, or which candidates to organize → The browser completes the work and summarizes → Close the browser after the sandbox → The company and HR confirm.
Keep daily posting for the same account, and recruiting for the same role, in the same session. Create a new session when the account, role, or rules change.
Matching human system
This matches an in-house LinkedIn channel executor: people confirm the role requirements and the final copy; posting, collecting applications, organizing candidates, and early outreach go to the AI. Interviews, offers, and onboarding stay with HR.
Steps
Open AI Employee → Enter market, find Independent LinkedIn Manager (linkedin01), and click Use.
After you enter that employee, click ⋯ → Enterprise Docs in the top right. Click the folder icon, New folder, name it LinkedIn Admin, then click Create.
Open the LinkedIn Admin folder (the top should show
LinkedIn Admin, not the tenant root). Only then click cloud upload and choose New Markdown document. The memory file must sit in this folder. On the first line write #Independent LinkedIn Manager memory (the memory name follows the employee: “[name] memory”). Click Save in the top right, renamedocument.mdto LinkedIn Admin.md, then Save again.In File info, fill in Title, Description, and Tags as needed, then click Confirm.
Click ⋯ on the LinkedIn Admin.md you just created, choose View info, and copy the url.
Click ⋯ → New session in the top right. Session name defaults to the name you just used; rename it to your own. Paste the copied url into Session prompt, and write: this is your memory, you fully own it, and you update your own important memory. Click Confirm.
Sample:
This is your memory:
(paste the url)
You fully own this memory. Update your important memory yourself.
Read this document before each job. Brand, posting rules, and job requirements are already in memory. Do not ask for them every time.LinkedIn needs a login. If the current session does not already have a logged-in account: it opens LinkedIn in a headed browser and gives you the browser address. Log in on that page, close it once, then open the browser again to verify the login. Log in once. Do not tell the AI the account or password. Next time it signs in by itself.
In the chat box, say what to post, which job to publish, or which candidates to organize or screen. To add materials, click the paperclip, open Add attachment, and choose Audio, Image, Camera, or File. Send.
If the first job or post is not right, keep saying how to change it in the same session, then ask it to publish or continue organizing candidates.
After each sandbox use, always close the browser.
Use different sessions for different roles or different kinds of LinkedIn work.