ChatGPT prompts for networking on LinkedIn
Anna SheronovaUpdated
The networking prompts worth pasting into ChatGPT or Claude are the ones that run on your own LinkedIn data, such as your message history and the people who viewed your profile. Each prompt on this page names the data it reads, either from your LinkedIn data archive or from a record of your activity connected to the assistant. The assistant ranks people and drafts messages for you to edit.
Two starter prompts, the week in review and people showing interest, need LinkedAsk, a Chrome extension that records your activity and connects to your assistant. The others run on files from your archive. Background on how an assistant reads LinkedIn data over a connection is in the LinkedIn MCP guide.
Which prompt needs which data
Your LinkedIn archive covers your connections, messages and invitations. Viewers and the reactions other people left on your posts are missing from it, so prompts about interest in you need a record kept as you use LinkedIn.
| Prompt | In your LinkedIn archive | In LinkedAsk |
|---|---|---|
| Who to follow up with | Every message with its date, sender, text and the profile URLs involved | Who wrote last and how many days ago, for people you saved |
| Week in review | Invitations alone, with no viewers and no reactions to your posts | Viewers, new followers, reactions, comments and invitations by day |
| People showing interest | Not in the archive | Viewers, followers and people who engaged with your posts, ranked |
| Warm introductions | Connections.csv has company and position for each connection | Your connections' headlines and your contact with each person |
| Network by company, role and month | Connections.csv, with company, position and the date you connected | Connections it has seen, and LinkedIn's connection total on each day it read it |
The archive sits under Me, Settings & Privacy and Data privacy, in the row Get a copy of your data. Pick the larger archive. The guide to LinkedIn's data export describes each file in it.
Prompt packs by job
Each pack lists the data every prompt reads and includes a fictional example answer.
- Who to follow up with on LinkedIn. Six prompts for replies you owe, stalled conversations and accepted invitations with no message since.
- Find warm introductions on LinkedIn. Five prompts that find people you know at a company and draft the request.
- LinkedIn network analysis with AI. Five prompts that chart your network, find who engages back and find who went quiet.
Four prompts to start with
Your LinkedIn week in review
This one reads profile viewers, new followers, reactions and comments on your posts, and invitations. It runs on LinkedAsk, since the archive has the invitations alone. With the archive, the accepted count comes from matching Invitations.csv against Connections.csv.
Review my last seven days on LinkedIn.
Tell me how many invitations I sent and how many were accepted.
List the people who viewed my profile, followed me, or reacted to or commented on my posts.
Name the people who appear in more than one of those lists.
If any data looks old or missing, say so instead of guessing.
People showing interest in you
Same lists as the weekly review, over 30 days, plus which people you already saved in LinkedAsk.
Find people who viewed my profile, followed me, or engaged with my posts in the last 30 days.
Skip anonymous viewers. Group the rest by person and rank the ones who appear more than once.
For the top five, tell me who they are, what they did, whether we are connected and whether I saved them.
Show me the list and I will decide whom to contact.
A brief before a call
For your messages with the person, the archive is the fuller source, since messages.csv has the full text. LinkedAsk adds profile visits in both directions and reactions, and holds message previews of about 50 characters for people you saved. Use both if you have them.
I have a call with [name] tomorrow. Tell me what my LinkedIn data shows about them:
when we connected, what our last messages were about, and whether they viewed my profile recently.
Then suggest two questions to ask, based on our last messages.
Reconnect with someone you lost touch with
messages.csv has the date and content of each message, which covers this one.
Using messages.csv, find my last conversation with [name] and summarize it in two lines.
Draft a short message to restart the conversation that refers to what we last talked about.
Keep it under 60 words, no sales pitch, and end with one easy question.
Two habits for prompts on LinkedIn data
Match people by profile URL. Names repeat and change, and both Connections.csv and messages.csv carry profile URLs. Tell the assistant to join files by URL and to ask you when a name matches two people.
Ask the assistant to report gaps. An empty answer from a recorded history may mean the tool never saw that activity. A line such as "if any data looks old or missing, say so" makes the assistant mention it before it draws a conclusion.
Run the prompts on your live LinkedIn history
LinkedAsk records your profile visits, viewers, followers, invitations and the reactions and comments on your posts as you use LinkedIn, plus message threads with people you save. You can connect it to your own AI so these prompts run on that record instead of a file. Claude supports custom connectors on all its plans. For ChatGPT, OpenAI lists full MCP connections for Business and Enterprise/Edu and read access for Pro, as of October 2026.
The connection includes four ready-made prompts, titled "Who should I follow up with today", "My LinkedIn week in review", "People showing interest in me" and "How my posts are doing". Adding it to your assistant is covered in connecting Claude to LinkedIn and connecting ChatGPT to LinkedIn.
Networking prompts: questions
What should I ask ChatGPT for LinkedIn networking?
Ask it to rank real people from your own history, for example conversations where the other person wrote last, or connections at a company you want to reach. Attach your LinkedIn archive or connect a record of your activity, and ask for a reason next to each name.
Can ChatGPT see my LinkedIn messages?
Only if you give it the data. Your LinkedIn archive includes messages.csv, which holds every message with its sender, date, text and the profile URLs involved, archived conversations included. Upload that file and the assistant can read it.
Do these prompts work in Claude as well?
Yes, they are plain text. To run them on a record of your activity instead of a file, the setup guides for Claude and ChatGPT show how to add the connection and which plans support it.
Which LinkedIn files do these prompts use?
Mostly Connections.csv and messages.csv from the larger data archive. Connections.csv has company, position and the date you connected for each 1st-degree connection. Invitations.csv records invitations, and you match it against Connections.csv to see which became connections.
Keep reading
- Guide · 5 minLinkedIn MCP: what it is and what it can readA LinkedIn MCP lets Claude or ChatGPT use LinkedIn data. Six kinds side by side by data source, what each reads and does, and where LinkedIn's ChatGPT app fits.
- Guide · 3 minFind warm introductions on LinkedIn with AIPrompts that find who you know at a company, rank them by your last message, and draft the request, with the LinkedIn data each prompt reads.
- Guide · 4 minWho to follow up with on LinkedIn: AI promptsPrompts that rank who to message on LinkedIn today, from replies you owe to profile viewers worth a hello. Each one says what data it needs and where to get it.
Give your prompts your LinkedIn history.
Add LinkedAsk to Chrome and connect it to Claude or ChatGPT. From then on the prompts can read your visits, viewers and invitations.