AI Cold email and linkedin system setup

JSK Advisory is a Sydney marketing agency doing local SEO and websites for service businesses.

The system has to run inside my existing Claude Code workspace and I use GHL so preferably link with this.

Scope
1. Domains and inboxes

3 to 5 secondary domains, 2 to 3 inboxes each, with SPF, DKIM and DMARC set up
Warm up and ramp each inbox to about 30 sends a day
2. Deliverability tracking

Daily Python collector into my database: sends, bounces, replies, unsubscribes, spam complaints, DNS health, blacklists, Google Postmaster reputation
Auto-pause an inbox and alert me in Slack when bounce or complaint thresholds are crossed
3. Lead generation

Pull local businesses by category and suburb from Google Maps data (I have DataForSEO)
Find decision-maker email s, verify every address, score leads against an editable YAML profile, and de-dupe against GoHighLevel
Claude writes a personalised first line per lead
4. Sending and replies

Campaigns run through Instantly or Smartlead via API
Claude classifies replies; unsubscribes go to a global suppression list straight away
5. Nurture

Qualified leads go into GoHighLevel with tags, a pipeline opportunity and a Slack alert
"Not now" leads go into a GoHighLevel nurture sequence of about 10 email s over 4 to 6 months
6. Claude commands

/outbound-leads, /outbound-review, /outbound-health and a weekly report
Must comply with the Australian Spam Act 2003 (sender ID, working unsubscribe, inferred consent).

You'll need
Proven cold email setups at volume (show bounce and reply rates)
Python, REST APIs and SQLite
Instantly or Smartlead API
GoHighLevel workflows and API
Real project experience with Claude Code or a similar coding agent
To apply, answer these
One cold email system you've built: daily volume, bounce rate, reply rate, tools used
The bounce and spam complaint rates at which you'd pause an inbox
Instantly or Smartlead for this, and why
Fixed price, timeline and monthly tool costs