- THE PLAYBOOK
- Posts
- How I’d build a predictable AI sales system on LinkedIn
How I’d build a predictable AI sales system on LinkedIn
In just 10 days:
The simple outbound system I’d build in 10 days
Most founders overcomplicate outbound on LinkedIn.
They try Clay, n8n, logic branches, and a dozen automations that break every other week.
After testing every setup you can imagine across 100+ companies, I’ve learned something:
Outbound only works when you get in front of the right buyer at the right time.
Not when you add more tools.
Here’s the exact 10-day blueprint I’d use to build a predictable AI sales system from scratch:
1. Build Phase
Most outbound fails because the foundation is missing.
Here’s what I lock in before sending message one:
Define the ICP
Interview top clients
Identify the business issues they were really paying to solve
Pull the themes you will use in your outreach
Select one clean data source (Sales Navigator is enough)
If your ICP is extremely niche, I build lookalike lists using Ocean.io.
This creates a clean seed list that will drive the entire system.
The goal for this phase is simple: clarity.
If your strategy is vague, your outbound dies before it starts.
2. Pick Your Sending App and Train the AI
I always use an AI SDR inside the executive profile.
It performs better because the message carries authority.
To train it, I load:
Dos and don’ts for tone
Positioning
Pains, benefits, differentiators
Past high-performing messages
When leads are uploaded, the AI scores them. I keep the high and medium scores and build a list of about 500 dream accounts.
If I need ultra-specific filters, I add Clay.
Predictability comes from clarity, not complexity.
3. Write the Sales Copy
Outbound dies when messages are long, vague, or generic.
I build two or three talk tracks:
A problem-led opener
A proof-driven angle
A curiosity gap that earns the reply
I keep every message at text length and run three or four follow-ups.
Tools I use: ChatGPT 5 and Twain.
Think of this like building a small library of plays you can run again and again.
4. Launch the System
You should be live by day 10. At this stage, I aim for:
15% or higher response rate on Sales Navigator
30% or higher on connection campaigns
If these numbers hit, you’ve proven message market fit.
The system is ready to scale.
5. Scale the Pipeline
By day 30, I double down on the campaigns that book meetings.
Most founders now have one or two predictable winners.
This is the point where outbound stops feeling random and starts feeling systematic.
From here, you can grow the system by adding more seats, more profiles, or routing conversations to a small team.
6. Handoff to a VA or SDR
By day 90, this becomes a 30-minute daily system.
You maintain:
Lead flow
Message library
Daily pipeline hygiene
Your team handles the rest.
Outbound becomes consistent and compounding.
Why This Works
Outbound breaks when founders chase complexity.
What scales is a simple engine built around:
Intent
Buyer readiness
Daily repetition
Signal-based messaging
This is how founders hit five to ten SQLs a month without burning their calendar.
You don’t need a $2,000 per month automation stack.
You need a clean system and a clear message.
Ready to go deeper?
I’m hosting a free 60-minute LinkedIn masterclass on December 12.
I’ll walk through the full outbound engine I use to help founders generate predictable pipeline with their personal profile and AI.
»» If you want to join, you can register for free here: Atticus Webinar
Only a few seats left.
See you inside,
— Luke @ Atticus