Cold outreach that does not read like a template
Bring prospect context, prior interactions, and message drafts into the same assistant conversation. Start with a small audience, review the reason for each message, and approve the exact recipients and cadence. Check queued and completed actions separately before deciding what to follow up.
Who this is for: Agencies starting conversations with potential clients and founders testing a specific offer.
Why this is worth automating
- Your offer matters more than a personalized opening line. Give the assistant approved claims and a clear reason this audience might care.
- Per-person drafts let you remove unsupported observations, false familiarity, and irrelevant asks before the message reaches someone else.
- Sequence and inbox tools let you inspect the next step in context. Keep an operator responsible for replies, opt-outs, and changes in the relationship.
How to run it, step by step
Use these prompts in Claude, ChatGPT, Instinct, or Meta Muse after connecting LinkedBoost. Review tool requests that change data or act on LinkedIn.
- 1
“Check our history with these five prospects and their companies. Flag existing conversations and current sequence membership.”
Exclude prior opt-outs and resolve overlap before preparing a new first touch.
- 2
“Create a draft sequence from my account with an invitation and one follow-up after acceptance. Do not enroll anyone.”
Review sender, timing, and text as a plan. Creating a sequence is separate from enrolling recipients.
- 3
“Rewrite Maya's opener using only the company context I supplied. Remove the implied claim that we have met.”
Keep edits specific to the person and within the note length available for the account and action.
- 4
“Show the exact person IDs and final drafts for approval. After I approve enrollment, report what is queued and what has completed.”
Enrollment can start scheduled outreach. Inspect returned state before retrying any action.
Plan your first run
| Checkpoint | Starting point | What to check |
|---|---|---|
| Suggested review batch | 5 people | Small enough to inspect every draft |
| Initial message | 1 reason | Explain a relevant reason to talk |
| Follow-up plan | Explicit | Choose cadence and stop conditions for this audience |
A practical example
Illustrative scenario: an agency offers onboarding research to software teams and has verified that a prospect owns onboarding.
Ask whether onboarding research is part of their current work. Say what we do without inventing a pain point, customer result, or familiarity. Show a draft for review only.
Example draft: 'Hi Maya, I run an agency that researches where new users get stuck during onboarding. Is that an area your team is working on this quarter?' Use it only when the role and offer are accurate. A reply, including a no, should guide the next step.
What goes wrong
Invented personalization
Do not reference a post the assistant has not read or claim familiarity you do not have. Remove facts without a source.
Retrying a queued action
A queued invitation is not yet a completed send. Inspect activity and sequence state before issuing it again.
Follow-ups outrunning the conversation
Check the thread and remaining steps after a reply. Stop further outreach when someone declines or asks not to be contacted.
Questions people actually ask
Is there a universal note length or cadence?
Use the character limit available for your account and action. Choose a concise message and an explicit follow-up plan; no universal cadence or note format guarantees replies.
Can I change one person's message?
The documented update_lead_messages tool supports per-person changes. Inspect current sequence state and final text before approving outreach.
Do caps make cold outreach safe?
No. Product limits do not make automation authorized by LinkedIn or eliminate restriction risk. Check account status and platform-policy guidance before using outreach tools.
Start with one account and one workflow
Connect LinkedBoost, inspect the account status, and run the first research or review step. Approve exact recipients and actions before outreach, then verify completion.
LinkedIn prohibits unauthorized automation. Account limits and operator review do not eliminate restriction risk. Read the LinkedIn User Agreement and our account-risk guide before acting. Native email, CRM sync, client dashboards, and white-label reports are not offered.
Related use cases
Build a prospect shortlist from your connected LinkedIn account or an authorized CSV. LinkedBoost gives your assistant tools to save lead groups, inspect prior interactions, and keep research alongside each person. Review who belongs on the list before drafting outreach or enrolling anyone.
Prepare a useful response to a specific LinkedIn post you have identified. LinkedBoost exposes tools for individual comments and reactions. Review the actual post, the account acting, and the proposed response before approving it. This workflow does not continuously monitor feeds or publish your own original posts.
Import an authorized CSV into a LinkedBoost lead group, then check which identities, fields, and history are available. Reconcile prior contact and exclusions before outreach. Importing leads does not guarantee that every message, reply, opt-out, or campaign state from the old system transfers.