Most LinkedIn outreach fails before the first message is sent. Not because the product is bad — but because the sender skipped the research and went straight to the template.
AI changes that equation. You can now do in 90 seconds what used to take 20 minutes: dig into a prospect's background, find a sharp angle, and write something that actually sounds like it was meant for them.
These 10 prompts are built for the full outreach workflow — from pre-send research to follow-up handling. Each one is ready to paste into ChatGPT, Claude, or any AI tool you already use.
(If you're running outreach through OutFlo, its AI-native engine already handles this selection for you — more on that at the end.)
Before You Reach Out: Research Prompts
Prompt 1 — The Prospect Deep-Dive
Use this before messaging anyone. Paste in whatever you have — LinkedIn URL, bio, recent posts — and let the model surface the angles worth using.
``` I'm about to send a LinkedIn connection request to a prospect. Here's everything I know about them:
Name: [NAME] Title: [TITLE] Company: [COMPANY] LinkedIn bio or summary: [PASTE BIO] Recent posts or activity: [PASTE ANY POSTS OR LEAVE BLANK]
My product/service: [ONE SENTENCE DESCRIPTION] My ICP: [WHO YOU TYPICALLY HELP]
Do the following: 1. Identify 3 things about this person that are directly relevant to why I'd be reaching out 2. Flag any trigger events (job change, funding, hiring spree, expansion) that make this a good time to reach out 3. Suggest the sharpest, most specific angle for a cold connection note — not generic flattery, something that would make them think "how did they know that?" ```
Prompt 2 — Company Signal Scanner
Before you message anyone at a company, you should know what's happening there. Use this prompt with a company description, recent news, or a LinkedIn company page summary.
``` I'm targeting [COMPANY NAME] for outreach. Here's what I know about them:
[PASTE: funding news, LinkedIn about section, recent posts, job listings, or anything from their website]
I sell: [WHAT YOU SELL] I typically help companies that: [DESCRIBE THE PAIN POINT OR USE CASE]
Based on the above: 1. What signals suggest this company is a good fit right now? 2. Are there any red flags or reasons this might be a bad time to reach out? 3. Which team or role at this company is most likely to care about what I'm selling, and why? 4. What's one specific, non-obvious thing I could reference in an outreach message that would make it feel researched? ```
Prompt 3 — Trigger Event Identifier
Timing is everything in outreach. This prompt helps you turn a raw piece of news into an outreach reason.
``` Here is a recent development about a prospect or their company:
[PASTE: LinkedIn post, job change notification, funding announcement, new product launch, hiring surge, etc.]
My offering: [WHAT YOU SELL]
Tell me: 1. Why does this event make them a relevant prospect right now? 2. What problem are they likely dealing with in the next 30–90 days that I could help with? 3. Write a one-sentence "reason for reaching out" I can use naturally in a message — referencing the event without sounding like I Googled them five minutes ago ```
Writing the Outreach: Message Prompts
Prompt 4 — Connection Note (Under 300 Characters)
LinkedIn's connection note limit is tight. Most people either waste it on "I'd love to connect" or write an essay. Neither works. Use this to write something that earns the accept.
``` Write a LinkedIn connection note for the following prospect. It must be under 300 characters, feel human, and give them a reason to accept — not sell them anything.
Prospect name: [NAME] Their role: [TITLE] at [COMPANY] What they're likely focused on: [1–2 SENTENCE CONTEXT] My angle / reason for reaching out: [WHAT MADE YOU REACH OUT] My role: [YOUR TITLE] at [YOUR COMPANY]
Rules:
- No "I came across your profile"
- No hollow compliments
- Must reference something specific, not generic
- Don't pitch. Just open the door.
```
Prompt 5 — First Message After Connecting
They accepted. Now what? This is where most people blow it. Use this to write a first message that starts a conversation rather than pitching into the void.
``` A LinkedIn prospect just accepted my connection request. Write a first message to send them.
Context:
- Their role: [TITLE] at [COMPANY]
- Why I originally reached out: [YOUR ANGLE OR TRIGGER]
- What I sell: [BRIEF DESCRIPTION]
- The problem I solve: [SPECIFIC PAIN POINT]
- Tone I want: [CONVERSATIONAL / DIRECT / CURIOUS — pick one]
Rules:
- Under 75 words
- End with a low-friction question, not a calendar link
- Do not mention demos, calls, or pricing
- Sounds like something a human would actually type, not a template
```
Prompt 6 — Hyper-Personalized Opening Line Generator
The first line of any cold message determines whether the rest gets read. Use this to generate a bank of openers for a specific prospect.
``` Generate 5 personalized opening lines for a LinkedIn message to this prospect.
Prospect info: [PASTE: anything relevant — bio, recent post, company news, job title change, content they've shared]
My product is: [BRIEF DESCRIPTION]
Each opener should:
- Reference something specific and real about them or their company
- Not sound like a compliment leading into a pitch
- Be 1–2 sentences max
- Feel like something you'd say if you ran into them at a conference and already knew who they were
Label each one A through E. ```
Prompt 7 — Follow-Up After No Reply
They didn't respond. Write a follow-up that adds value instead of just bumping the thread.
``` I sent a LinkedIn message to a prospect 5–7 days ago and got no reply. Write a follow-up message.
My original message was about: [BRIEF SUMMARY] Their role: [TITLE] at [COMPANY] The value I can offer them: [SPECIFIC OUTCOME OR INSIGHT]
Rules for the follow-up:
- Do not open with "Just following up" or "Bumping this"
- Lead with something new: a useful insight, a relevant stat, a short observation about their space
- Keep it under 60 words
- End with a yes/no or one-word question — the lowest possible friction
- Make it easy to respond even if they're not ready to buy
```
Prompt 8 — Handling the "Not Right Now" Reply
Someone replied but pumped the brakes. This is not a dead lead — it's an opening. Use this to write a response that keeps the relationship alive without being pushy.
``` A prospect replied to my LinkedIn outreach with something like: "[PASTE THEIR REPLY]"
My product: [WHAT YOU SELL] The problem I solve: [PAIN POINT]
Write a reply that: 1. Acknowledges what they said without being dismissive or aggressive 2. Offers something low-commitment — a resource, a thought, a relevant question 3. Leaves the door clearly open without pressuring them 4. Is under 80 words and sounds completely human ```
Campaign and Strategy Prompts
Prompt 9 — Full Outreach Sequence Builder
Use this to build a multi-touch LinkedIn sequence from scratch for a specific ICP segment.
``` Build a 4-step LinkedIn outreach sequence for the following target:
ICP: [DESCRIBE: role, company type, size, stage] My product: [WHAT YOU SELL] Core pain point I solve: [BE SPECIFIC] Desired outcome of the sequence: [BOOKED CALL / DEMO / REPLY / AWARENESS]
For each step, provide:
- Timing (e.g., Day 0, Day 5, Day 10, Day 18)
- Action (connection request / message / InMail / comment on their post)
- Full message copy
- The goal of that specific touch
Rules:
- Step 1 must be a connection note that doesn't pitch
- At least one step should lead with value (insight, resource, stat) instead of asking for something
- Final step should be a clean "closing the loop" message, not a guilt trip
```
Prompt 10 — ICP Refinement from Replies
This is the most underused prompt on the list. After any campaign runs, feed your reply data back into this to sharpen who you should be targeting.
``` Here are replies I've received from a recent LinkedIn outreach campaign:
Positive replies (interested or engaged): [PASTE 3–5 REPLIES]
Neutral or deflecting replies: [PASTE 2–3 REPLIES]
No replies — notable patterns I noticed: [DESCRIBE: e.g., "mostly VP-level ignored me, manager-level replied more"]
My current ICP assumption: [DESCRIBE WHO YOU WERE TARGETING]
Based on this, help me: 1. Identify what the positive responders have in common (role, language, company stage, pain signals) 2. Identify where my ICP assumption was wrong or too broad 3. Recommend 2–3 adjustments to who I should be targeting or how I should be framing the outreach going forward ```
What OutFlo Does Instead of You
Every prompt above solves a different problem at a different stage. Connection note, opener, follow-up, objection reply, closing loop — each one requires a different approach, and choosing the wrong one at the wrong moment is usually what kills a deal quietly.
OutFlo's AI-native engine removes that decision layer entirely.
It reads exactly where you are in a campaign — connection sent, first message delivered, follow-up overdue, soft reply received — and selects the right message approach automatically. Not a generic template. The right message, calibrated to the prospect, the sequence step, and the momentum (or lack of it) in that specific conversation.
The follow-up after three days of silence looks different from the follow-up after a "maybe later." OutFlo knows the difference and writes accordingly, step by step, across every active lead in your pipeline — so no deal slips through because the wrong message went out at the wrong time.
The 10 prompts in this post are the manual version of what OutFlo runs natively. Use them to understand the logic. Then let OutFlo run the logic at scale.
How to Get More from These Prompts
The prompts above work as-is, but they get sharper with better inputs. A few principles:
More context = better output. Don't just put a job title. Paste the full LinkedIn bio, a recent post, a funding announcement. The more specific the input, the more specific the output.
Iterate, don't accept the first draft. Use these as starting points. Tell the AI what's off: "too formal," "sounds like a template," "make it shorter." Three rounds of refinement beats one lucky first draft.
Build a swipe file. When a prompt produces something great, save it. The goal over time is a library of angles and openers that work for your specific ICP — not a fresh prompt every time.
Personalization doesn't replace relevance. Knowing someone's job title isn't personalization. Knowing they just expanded to a new market, or posted about the exact problem you solve, is. Let the research prompts do the work before you get to the message prompts.
The best outreach doesn't feel like outreach. These prompts won't do the thinking for you — but they'll cut the time between "I should reach out to this person" and "I actually sent something worth reading" from an hour to five minutes.
That's the compounding advantage.
