3 ChatGPT & Claude Prompts for a LinkedIn Connection Request

Copy three ChatGPT and Claude prompts for a LinkedIn connection request, with example drafts for an event, a post, or a specific question. Check every fact before sending.

The note on a connection request has room for one real reason to connect. ChatGPT or Claude can help cut the wording, but neither knows why you picked this person unless you tell it. Feed it a profile and ask for a “personalized outreach message,” and you’ll usually get a polite sentence that could go to fifty people.

Choose a prompt for the reason you actually have:

• An event you attended.
• A post you read.
• A specific question about their work.

The prompts below aren’t finished notes. They’re small editing jobs. Supply one true detail, ask for a short draft, then check that the result says no more than you know. If there isn’t a real reason to write, skip the prompt and the invitation. When a blank request beats a weak note: Should you add a note to this LinkedIn connection request?.

First, know how much space you have

Checked October 31, 2026: LinkedIn’s current invitation guide says a basic account can add notes to up to three connection requests per month, each up to 200 characters. Another LinkedIn help page says five personalized invitations for basic accounts, also at 200 characters; it says Premium members can send unlimited personalized invitations with 300-character notes. LinkedIn’s own pages disagree on the free monthly count. The two pages side by side, plus what 200 characters actually fits: the character limit, explained. Check what your account shows instead of planning around either number. The prompts aim below 200 characters, with space to edit.

The monthly note cap is separate from the weekly invitation limit, which governs how many invitations you can send at all: the weekly limit, explained.

These are invitation notes, not InMail and not the message you send after someone accepts. LinkedIn’s AI Writing Assistant can also draft an invitation for Premium members using English. LinkedIn itself tells users to review and edit that draft. The same advice applies when the draft comes from ChatGPT or Claude.

Prompt 1: a shared event you actually attended

Give the model the event and a moment you remember. If you didn’t speak to the person, say so; “great meeting you” would be a lie.

Write one LinkedIn connection note under 180 characters. I attended [event] and heard [person's name] ask [specific question] during [session]. We did not meet. Mention that question and say why I'd like to stay in touch: [real reason]. No praise, pitch, or claim that we spoke. Use plain language.

Example input: Maya asked at the local revenue meetup how a small team tracks replies after an event. The sender also works on follow-up processes. A draft could be:

Maya, your question at the revenue meetup about tracking replies stuck with me. I'm working on the same follow-up problem. Would be good to connect.

Before sending, confirm the question was hers. If it came from the moderator, the sentence fails even though it sounds specific.

Prompt 2: a post that changed your thinking

A vague compliment is cheap. Tell the model which point in the post you want to remember. Don’t imply you tested their advice unless you did.

Draft one LinkedIn invitation note under 180 characters. [Name] posted [link or exact excerpt]. Their specific point was [one accurate point]. I want to connect because [true reason]. Mention the point, not how "insightful" they are. Don't invent a result, a relationship, or a question just to fill space. Give one option only.

Example input: Ravi wrote that his team stopped measuring only meeting bookings and started checking whether first replies led to useful conversations. The sender is changing their own reporting. A draft:

Ravi, your post on looking past meeting counts made me rethink our outreach report. I'm changing what we track too. Would be good to connect.

Read the original post before using the line. If Ravi said the opposite, a smooth sentence won’t rescue it.

Prompt 3: a specific question about their work

Sometimes the connection is a question, not shared history. Keep it small enough to answer later. A connection request isn’t a survey or a meeting ask.

Write one LinkedIn connection note under 180 characters. [Name] works on [verified role or project]. I'm trying to understand [narrow question]. Ask whether they're open to connecting around that topic. Don't say we know each other, don't praise their career, and don't ask for a call. If the role or project is missing, ask me for it instead of guessing.

Example input: Jo leads onboarding at a company that recently added a self-serve trial. The sender wants to understand how her team handles trial handoffs. A draft:

Jo, saw your team added a self-serve trial. I'm curious how you handle handoffs from trial to onboarding. Open to connecting?

Edit the output, not the facts

Run the draft through four checks before it becomes a note:

1. Truth: Can you point to the event, post or project it names? Does it invent a conversation or imply you used their product?
2. Purpose: Would this still make sense if you removed the person’s name? If yes, it needs a more specific reason or it should not be sent.
3. Voice: Delete stock openings like “I came across your impressive profile” and sales language like “explore synergies.” A normal sentence is enough.
4. Length: Count the final text after replacing names and details. Don’t trust the model’s character count; paste the note into LinkedIn and check the field before sending.

If the output misses a check, edit it yourself or ask for a constrained rewrite:

Keep the factual detail about [verified detail]. Remove the generic compliment and any claim we know each other. Shorten this to under 180 characters. Do not add new facts: [paste draft]

The same constrained prompting, tested on follow-ups after silence: LinkedIn follow-up after no response: we tested ChatGPT and Claude prompts.

For a worked check of those boundaries, our LinkedIn DM editing test compares three fictional drafts and the actual ChatGPT edits against the supplied facts.

The same check applies when the AI touches your lead list, not just your note: how to check an AI-cleaned spreadsheet before you trust it.

Fifteen founder outreach templates cover more situations when you already know what to say. For the inbox side, see why LinkedIn messages go unanswered.

Checked October 31, 2026: LinkedIn’s invitation guide, its InMail-versus-invitation page and AI Writing Assistant help (linked above). LinkedIn’s two help pages conflict on the basic-account monthly personalized-invitation count. All people, events and companies in the example notes are fictional; the notes are illustrations, not tested response-rate claims.