Use a reviewed template when the people in your audience share a relevant problem but you have little reliable context about each person. Use a reviewed AI-written draft when a verified detail changes what is worth saying to that prospect. If the detail is shaky, a specific-sounding message is not an upgrade: return to the template or wait until you have a sound reason to reach out. This is a decision about producing campaign messages, not whether a connection invitation needs a note.
Choose by the context you actually have
Before drafting, write down the audience, the shared reason to contact them, and any individual detail you would rely on. Then ask whether that detail would change the message’s premise or question. A name, job title or generic compliment rarely does. A prospect’s own account of a problem might—provided you can check it and explain why it connects to what you offer.
The trade-off is review effort, not a proven reply-rate advantage. A template asks you to validate one shared premise and check that each recipient fits it. A prospect-specific draft asks you to validate the premise and each personal reference. That extra work can be worthwhile for a small, well-researched audience; across a thinly researched list, it can turn plausible wording into inaccurate claims.
| Context available | Action | Reason |
|---|---|---|
| A shared, relevant problem; few dependable individual details | Use a reviewed template | The common premise does the work without invented specificity. |
| A verified detail changes the useful question or angle | Draft for that prospect, then review it | The reference gives the recipient a more relevant reason to respond. |
| Neither a sound shared premise nor a trustworthy individual reason | Hold the message | Neither writing method can fix a poor-fit approach. |
Compare both messages for the same audience
Consider a fictional sender who helps small software companies clarify sales-to-onboarding handoffs. Their audience is operations leaders at those companies. The following reusable message assumes only that handoffs are relevant to the role; the sender would still check that each company and recipient fit before using it: “I work with operations leaders at small software companies on sales-to-onboarding handoffs. Is clarifying who owns the next step something your team is working on?”
Now suppose a fictional prospect, Morgan, has published a post saying that ownership gets unclear when a customer moves from sales to onboarding. A prospect-specific draft could say: “Morgan, your post about unclear ownership between sales and onboarding caught my attention. I help teams clarify that handoff. Is identifying the next owner the part your team is trying to improve?” The difference is not simply Morgan’s name. The question follows a stated issue rather than asking about a possible issue across the whole audience.
Before sending the second version, open the post and confirm Morgan wrote it, that it concerns their team rather than an example about another company, and that the draft describes it accurately. If you cannot verify those points, remove the reference and use the reviewed template if Morgan still fits the audience. Do not turn a public observation into a claim that you know the team’s internal process.
Review what the recipient will actually read
Review a template against the audience before using it, then check that each intended recipient still fits. Review a personalized draft one recipient at a time. The fields below provide a practical pre-send check; if a reference fails, delete it rather than hoping the recipient will overlook it.
- Identity: Is this the right person, company and role? Could two people with similar names have been confused?
- Accuracy: Can you locate the post or other detail the message refers to? Does the draft preserve what the person actually said?
- Relevance: Does the detail change the reason for writing or the question? If not, keep the simpler wording.
- Implied knowledge: Does the message suggest you know about internal priorities, results or problems that the prospect has not disclosed?
- Workflow: If a person must approve each prospect-specific reference, use a review-first sending process rather than treating an unattended send as approved.
Match the method to your plan and risk tolerance
In ZenMode, templates go out as written on every plan, with fields such as first name and company filled in for each prospect. On Practitioner and higher plans, the AI can draft your template for you to edit, or, in "AI writes every message" mode, write each prospect's sequence from their name, role, company and location. It does not read their posts, so a reference like the one to Morgan's post has to come from you. In that mode you can have each sequence wait for you to approve or edit it before it goes, rather than letting Autopilot send it. Do that for any campaign that relies on checked prospect details. No plan choice makes an unverified personal reference accurate, so check it before it goes.
Account risk is a separate choice from message quality. LinkedIn says it does not allow third-party software or browser extensions that automate activity on its website. Reviewing better copy does not change that policy. If that risk is unacceptable for your account, write and send messages manually, whether you start from a template or a prospect-specific draft.
