By the end of this tutorial, you’ll know how to add an AI-assisted Comment last post step to a lemlist sequence, configure how the comment is generated, and review the task before publishing it on LinkedIn.
This workflow helps you scale thoughtful LinkedIn engagement without losing control. Instead of writing every comment from scratch, lemlist drafts a contextual comment based on your prospect’s latest post, then creates a manual task so you can review, edit, and publish it yourself.
Why this matters
Commenting on a prospect’s content is one of the best ways to get noticed before sending a direct outreach message. A relevant comment puts your name, profile, and perspective in front of the prospect in a more natural way than an invite or cold message.
With AI-generated drafts and manual approval, you can add this high-impact touchpoint to more sequences while keeping your comments authentic, on-brand, and safer from spammy behavior.
Prerequisites
You should already know how to create or edit a sequence in lemlist.
You should already have a LinkedIn account connected as a sender.
You should already understand how manual tasks work in your workflow.
Core lesson: Configure the AI Comment last post step
Phase 1: Add the LinkedIn comment step to your workflow
In the Sequence tab, click Add step, open Manual steps, and select Comment last post.
This adds a manual LinkedIn step that reads the prospect’s latest post, drafts a relevant comment with AI, and sends it to your task list for approval instead of posting automatically.
Phase 2: Choose who will post the comment
After adding the step, choose the LinkedIn account used to comment on last post.
This tells lemlist which connected sender should be used when the task is eventually approved and completed. If your workspace uses multiple senders, selecting the right one keeps ownership and activity aligned with the right profile.
Phase 3: Control the style of the AI-generated comment
Set the Comment tone that best fits your outreach style.
This is how you guide the AI’s voice. Depending on your workspace, you may see options such as Professional, Friendly, Casual, Enthusiastic, or Thoughtful. Choose a tone that matches your brand and the type of relationship you want to build with prospects.
Choose the Comment length: Short, Medium, or Long.
This controls how concise or detailed the draft will be. Short comments work well when you want a lightweight touchpoint, while medium or long comments can add more perspective when the post deserves a more substantial response.
Use Guidelines for AI comment generation to add custom instructions.
This is where you can steer the draft beyond tone and length. For example, you can ask the AI to mention a specific angle, avoid sounding too promotional, ask a follow-up question, or keep the wording aligned with your company voice.
Phase 4: Define when the step should apply
Set Max post age (days) and Skip if not approved (days).
Max post age tells lemlist to skip the step if the prospect’s latest LinkedIn post is older than your limit. This helps keep comments timely and relevant. Skip if not approved automatically skips the manual task after a set number of days so your sequence can continue moving if no one reviews it in time.
Phase 5: Review and publish the comment from Tasks
When the step runs, lemlist creates a task in your Tasks list with an AI-generated comment draft.
Review the draft carefully, make any edits you want, and click Do to publish the comment on LinkedIn. This final review is what keeps you in control of quality, tone, and relevance.
You stay in control. The AI suggests the comment, but nothing is posted automatically. Every comment is manually reviewed before it goes live.
Best practices for better AI-generated comments
Match the tone to the audience: Use professional for formal industries, casual for founder-led or creator-style outreach, and enthusiastic when you want more energy.
Keep custom guidelines specific: “Reference the main insight and ask a short question” will usually perform better than broad instructions like “make it better.”
Use a realistic post age limit: Commenting on very old posts can feel awkward or inauthentic.
Set an auto-skip window: If your team reviews tasks in batches, this prevents stalled leads from blocking the next steps in the sequence.
Edit before posting: Even strong AI drafts should be checked for nuance, brand fit, and context.
Practical application: A simple real-world setup
Here’s an example of how a sales team might use this step in practice:
Tone: Professional
Length: Medium
Guidelines: “Acknowledge the main takeaway from the post, sound thoughtful, and end with a light question when relevant.”
Max post age: 14 to 30 days
Skip if not approved: 3 to 7 days
This setup works well when you want comments to feel timely and personalized, but still efficient enough to scale across a large prospect list.
Troubleshooting and common pitfalls
Issue: No comment task is created
Root cause: The prospect may not have a recent LinkedIn post that fits your max age setting.
Fix:
Increase the Max post age value.
Check that the lead has an accessible LinkedIn profile and recent activity.
Test the step on a small group of active LinkedIn users first.
Issue: The draft feels too generic
Root cause: The prompt is too broad or the selected tone/length is not restrictive enough.
Fix:
Add more precise Guidelines for AI comment generation.
Ask the AI to reference the post’s main idea, insight, or opinion.
Review whether a shorter or more thoughtful tone would sound more natural.
Issue: The sequence stops waiting for approval
Root cause: The task has not been approved yet, and the skip window is too long or not aligned with your team workflow.
Fix:
Reduce the Skip if not approved setting.
Review tasks more frequently.
Use a shorter approval window for high-volume sequences.
Issue: The wrong LinkedIn sender is being used
Root cause: The step is assigned to a different sender than expected.
Fix:
Double-check the selected sender in the step settings.
If needed, force a specific sender for that step.
Confirm the chosen LinkedIn account is connected and active.