Draft a personalized LinkedIn® message in three steps: write your one reason for contacting this person, give an AI assistant only facts you can verify about them and your relationship, and ask for one small next step they can easily accept or decline. Then edit the draft line by line and send it yourself.
Key takeaways
- A message is personalized when it explains why this conversation makes sense for this person now; a name and job title alone do not do that.
- Give AI a six-part brief (recipient context, relationship, reason, evidence, next step, boundaries) and tell it to ask you about missing facts.
- Size the request to the relationship: a first message asks a question or permission, while an existing relationship can carry a clear request.
- Review and send every message yourself, one at a time; drafting help gives you no permission to automate outreach.
- Follow up only with new information, and stop when someone declines or asks you to stop.
What makes a message personalized?
A personalized message explains why this conversation makes sense for this person now. It connects a relevant observation to a clear purpose. Mentioning a job title, company, or recent post is useful only when it helps the recipient understand that connection.
Personalization is the use of relevant, accurate context to shape the substance of a message. “I saw you work in operations” is a detail, but not a reason to talk. “Your post about missed renewal dates raised a question about how your team records account ownership” gives the recipient a specific subject to respond to.Avoid artificial intimacy. Say you found someone's profile today if that is what happened, and claim a shared interest only when you can name it. A candid introduction can be more credible than a forced connection: “I found your article while researching how teams run renewal reviews” explains the context without pretending a relationship already exists.
LinkedIn®'s Professional Community Policies set expectations for authentic and respectful conduct. The advice here on relevance and proportionate requests applies those principles to private messages.
How do you choose an intent before writing?
Choose one purpose: ask a question, request an introduction, discuss a role, offer a relevant resource, continue an existing conversation, or explore a business fit. Combining several purposes makes the message harder to answer and can disguise the real request.
Write the purpose in a sentence before opening an AI assistant. For example, “I want to ask whether this person is open to sharing how they run renewal reviews.” That purpose calls for a focused question. It does not justify a calendar link, an attached presentation, and a request for a referral in the same message.
Consider the recipient's likely effort. Asking someone to confirm whether a resource is relevant is different from asking them to review your strategy. A short message can still contain a large request. Make the expected work visible so the person can decide freely.
| Relationship | Reasonable starting point | What needs more context |
|---|---|---|
| No previous interaction | A specific question or permission to share something relevant | A meeting request with no explanation |
| A recent public conversation | Continue the exact topic already discussed | A sudden unrelated sales pitch |
| An established working relationship | A clear request tied to shared work | Assuming availability or agreement |
Use the table as judgment guidance, not a rigid etiquette rule. Some professional contexts welcome direct commercial introductions. The requirement is honesty about your purpose and enough information for the recipient to decide whether it is relevant.
What context should you give the AI assistant?
Provide the recipient context you actually know, the relationship, your purpose, your relevant evidence, and the desired next step. Separate these from assumptions. A small structured brief is more useful than a large pasted profile with no explanation of why you want to talk.
Copy this brief and fill in the brackets before you open the assistant:
- Recipient context: [the public work or prior conversation relevant to my request]
- Relationship: [how we know each other, or “none yet”]
- My reason for contact: [the specific question or offer]
- My evidence: [facts I can support]
- Requested next step: [one action]
- Boundaries: no invented familiarity, no unsupported outcomes, no pressure language
If you cannot fill the reason or the next step, the message is not ready to draft.
OpenAI's prompting guidance recommends being specific and refining the response. For messages, the refinement should focus on relevance and clarity. “Make this more persuasive” can produce inflated language; “make the reason for contact clearer without adding claims” is a more useful instruction.
Use only the information needed to draft. Remove private contact details, confidential correspondence, and sensitive personal information. Review the provider's data practices and your organization's rules. For ChatGPT, OpenAI's data controls distinguish model-training settings from retention and other account controls, so changing one setting can leave other privacy requirements open. Check each one that applies to you.
What is a reusable drafting prompt?
A good prompt defines the task and the evidence boundary. It tells the assistant to ask you about missing context instead of filling every gap with plausible language.
Copyable prompt: “Draft a LinkedIn® message from the brief below. Explain the specific reason for contact, use only verified context, and make one proportionate request. Keep the tone respectful and direct. Use only the relationship, reading and results in the brief, and add no testimonials. If a fact is missing, ask me a question instead. Return two versions and explain the difference in the requested effort.”After the draft, use a second prompt: “Identify any sentence that assumes something about the recipient, overstates my evidence, or asks for more work than the relationship supports. Suggest a narrower version.” This turns the assistant into an editor that narrows the pitch instead of making it more forceful.
Keep your own words where they are already clear. AI may replace a simple question with a formal introduction or unnecessary praise. Reject those changes and keep your original line. Google's clear-writing guidance favors specific subjects and verbs, a useful principle when a message must communicate its purpose quickly.
Which message templates can you adapt?
The following examples are illustrative, with invented names and situations, and report no real outreach results. Use them only when their underlying relationship and facts are true. Replace the substance before changing the tone. A polished template with an irrelevant offer remains an irrelevant message.
Template 1: Continue a public discussion
“Hi Maya, your explanation of renewal reviews answered my question about account ownership. I am still thinking about the exception you mentioned when two teams share an account. Do you record that as joint ownership or name one final decision-maker? Happy to keep the discussion in the public thread if that is easier.”
Use this after a real exchange. It references the exact point and gives the recipient a choice of channel. A public reply invites more discussion of that topic, so keep sales out of the follow-up. If you have no new question, a brief thank-you may be enough. The purpose is to continue an existing subject, not manufacture a reason to enter someone's inbox.
Template 2: Ask permission to share a resource
“Hi Jordan, your post about onboarding documentation raised the same distinction our team is working through: instructions for setup versus instructions for recovery. I have a short checklist on that distinction. Would it be useful if I sent it over?”
This works only if the resource exists and matches the problem; an unrelated brochure turns the permission question into a pretext. If the resource is publicly accessible and the context makes a direct link natural, a permission step may be unnecessary. Choose the version that reduces friction for the recipient, not the version that creates an extra reply to count.
Template 3: Make a transparent business introduction
“Hi Alex, I help software teams set up renewal reviews before contracts come due. Your article about late renewal conversations made me wonder whether that is a current priority for your team. If it is, I can explain the kind of scoping work I do; if not, no need to reply.”
This is a commercial message and says so openly. Adapt the offer to your actual service and remove claims you cannot support. A clear introduction may still receive no answer; the recipient may simply have different priorities.
Template 4: Ask about a role
“Hi Taylor, I am considering the implementation specialist role on your team. The description emphasizes customer training and account reviews. My background is in customer education, and I would like to understand how much of the role involves configuration work. Is there a public resource or recruiting contact you would recommend?”
Use this when you have read the role description and have a question it leaves open. Check the careers page first, so you never ask an employee to repeat it. Keep your experience accurate and make it easy to redirect you. The recipient may not be involved in hiring and should not be treated as an unofficial recruiter.
Template 5: Reconnect with a former colleague
“Hi Casey, I was reviewing the documentation from our migration project and remembered how useful your launch checklist was. I am now working on a similar coordination problem. If you are open to it, I would enjoy hearing how your approach has changed since then. No rush if your schedule is full.”
Use a memory that both people can recognize, and leave out nostalgia you do not share. If you need a specific favor, state it plainly instead of hiding it behind an open-ended catch-up. An established relationship gives you context, but does not create an obligation to provide free consulting, introductions, or referrals.
Template 6: Follow up with a useful update
“Hi Morgan, a quick update on the question I sent about onboarding reviews: I found a public guide that answered the setup part. The remaining question is who owns the final approval when delivery spans teams. If that is outside your area, no problem; I will keep researching.”
Use this when the update genuinely changes the conversation. “Just checking in” offers no new information and can sound like a demand for attention. A follow-up should make the request easier to assess. Before you send one, check the table:
| Situation | What to do |
|---|---|
| You have new, relevant information | Send one short update that says what changed |
| No reply and nothing new to add | Wait, or leave the conversation there |
| The person said “not now” | Note the context they gave and decide later whether it still applies |
| The person declined or asked you to stop | Stop |
Template 7: Close a conversation respectfully
“Thanks for letting me know. I understand this is not a priority right now. I will leave it there, and I appreciate the clear answer.”
A closing message does not need another objection-handling attempt. If the recipient invites future contact, note the context they provided and decide later whether it remains relevant. “Not now” gives you no permission for a recurring reminder sequence. The relationship is more important than completing every step in a sales template.
How do you edit a message from generic to relevant?
Consider this illustrative first draft: “Hi Sam, I have been following your amazing work and am impressed by your leadership. We help companies transform onboarding and achieve exceptional outcomes. Are you free for a quick call this week?” The problems are not merely stylistic. It invents familiarity, uses unsupported claims, and requests time before explaining the fit.
Start by checking the facts. Suppose the sender actually read one public article about unclear implementation ownership. They offer scoping workshops but have not supplied any measured customer results. A truthful opening is: “I read your article about unclear ownership during implementation.” It states exactly what happened.
Connect the observation to the offer: “That is the problem I focus on in scoping workshops.” This gives context without claiming the recipient has the problem today. Then narrow the request: “Is that a topic your team is currently working on?” The recipient can answer without committing to a meeting.
The revised message is: “Hi Sam, I read your article about unclear ownership during implementation. That is the problem I focus on in scoping workshops. Is it something your team is currently working on? If so, I can send a short outline of the decisions a workshop covers.” The message is still commercial, but the purpose and next step are visible.
Before sending, check whether the article is recent enough to support the reference and whether Sam's role is relevant. If not, change the context or leave the message unsent. Personalization is a judgment about fit, not a decorative sentence added after the audience has already been chosen.
Where does drafting assistance stop?
Drafting assistance prepares text, and the sending stays with you. LinkedIn®'s User Agreement restricts unauthorized automated methods, including certain account actions, and a tool's timing controls or suggested daily allowance give you no safe automation limit for connection requests, messages or follow-ups.
Reepl is an AI writing platform that puts your social media on autopilot in your own voice: it finds what to post about, drafts your posts, and keeps your calendar full, and you decide how much it does without asking. For messages, Reepl drafts wording for you to review, edit, and send. Reepl is not an outreach automation system.
Keep the review step concrete, and run it on every message:
- Read the whole draft aloud.
- Confirm the recipient and the account you are sending from.
- Compare every personal reference with what you actually know.
- Check that the request fits the relationship.
- Send it yourself, or leave it unsent.
A different first name in each message is not a different reason for contact. Each conversation needs its own reason.
If your task is a short introduction, the connection request writer can help structure that reason. It cannot determine whether the person wants to hear from you. Your responsibility includes deciding not to send when the context is weak.
What should you track after sending?
Track whether messages lead to relevant conversations and whether your commitments are fulfilled. A reply count alone does not distinguish interest from a correction, a decline, or a request to stop. Categorize responses in a way that reflects the actual purpose of your communication.
For an illustrative recruiter, a useful result might be a candidate clarifying location requirements. For a consultant, it might be a prospect explaining that the problem exists but belongs to another team. Neither outcome is necessarily a meeting, yet both can improve understanding. Avoid converting every response into a success claim.
Keep notes limited to legitimate professional context and follow your organization's retention requirements. Record what you promised, what the person requested, and whether further contact is welcome. Leave out unrelated personal details, even if they would make a later message sound more familiar.
If messages receive little response, review targeting and the size of the request before changing adjectives. Ask whether the offer is relevant, the evidence is specific, and the recipient can understand the next step. The recruiter workflow overview offers related context for professional communication, but no template can remove the recipient's right to decline.
What should you do before your next message?
Choose one real person and write the reason for contact in your own words. If you cannot explain the relevance without flattering language or assumptions, gather better context or leave the message unsent. Then ask AI to improve clarity while preserving that reason.
Review the resulting draft for truth, effort, and tone. Confirm that every reference to your relationship is true and that the recipient can respond without being cornered into a meeting. Send it yourself only when the wording represents what you actually mean.
A useful message is not the most elaborate one. It makes a legitimate purpose easy to understand and gives the other person room to decide. Keep that standard through the reply, follow-up and closing, and your drafting process will support the relationship.
Send Better LinkedIn® Messages with AI
Smart messaging assistant with intent-based responses for sales, recruiting, and networking, right inside LinkedIn®.
