AI Cold Email for SaaS Sales Teams
Founder of Link AI. Writes about voice agents, cold email, and the operational reality of running them in Hebrew and English.
SaaS outbound has a copy-paste problem. Templates with merge tags don't pass the smell test anymore — buyers can spot 'Hi {firstName}, I saw you raised funding' from the subject line. Mailer reads each lead like an SDR would, finds the actual reason to email them (a recent product launch, a job posting, a podcast appearance, a specific GitHub commit), and writes the email around that fact. The result is roughly 70-80% open rates and ~10% reply rates on cold campaigns where the ICP is tight.
Why SaaS outbound is broken
Most SaaS sales teams have tried the 'personalization at scale' playbook: Apollo + Lavender + a {firstName}{companyName} merge tag, blasted at 1,000 leads per week. The opens are flat, the replies are negative, and the team burns the domain.
The problem is not the volume. The problem is that buyers can pattern-match an outbound template in under three seconds and the template doesn't reference anything they actually did.
What Mailer does for SaaS sales teams
Mailer takes an ICP description in English (or Hebrew), finds real matches via Apollo + LinkedIn + signal sources (job postings, funding announcements, GitHub, podcast guest lists), and writes a per-lead email that references something concrete. Each email gets two follow-ups, also context-aware. Replies go through Mailer's reply-handling layer so positive replies route to a human and out-of-office goes back into the queue.
- ICP described in natural language, not 47 dropdown filters
- Per-lead research from Apollo, LinkedIn, and signal sources
- Original email body per lead — not template + merge tag
- 2 follow-ups, each aware of the prior email
- Reply handling that routes positives to humans and reschedules OOO
- Domain warmup and deliverability monitoring built in
How Mailer writes a SaaS cold email
Example ICP: 'Heads of Engineering at Series-A SaaS companies in fintech who are hiring backend engineers right now.' Mailer pulls the list (maybe 80 matches), enriches each with their GitHub recent commits, their team's job posting language, and their company's recent funding context, then writes 80 distinct emails — each referencing the specific job posting, the engineering culture signal in that posting, and Link AI's relevant capability.
Per Link AI's analysis of 50,000+ Mailer campaigns, the lift from per-lead research over template + merge tag is on the order of 3-5x reply rate, not 20-30%.
ROI math for a SaaS sales team
A 3-SDR team sending 500 outbound emails per SDR per week (1,500 total) at template-and-merge-tag quality gets 1-2% reply rates, maybe 15-30 conversations per week, of which 4-8 turn into meetings. Mailer's per-lead approach cuts volume to ~400 per SDR per week and lifts reply rates to 8-12%, which means 35-50 conversations and 10-18 meetings.
Frequently asked questions
How is this different from Apollo + Lavender + Outreach?
Apollo finds leads. Lavender critiques drafts. Outreach sends sequences. Mailer does the actual research and writing per lead — the part those tools assume an SDR will do. The lift is on the per-lead body, not on cadence or filters.
What about reply rates?
Typical Mailer campaigns land 7-12% reply rates on tight ICPs in SaaS. Loose ICPs (anyone with 'VP' in title at any SaaS company) drop to 3-5%.
How does deliverability work?
Mailer warms domains for 4-6 weeks before any outbound, monitors per-mailbox health, and rotates mailboxes automatically. The warmup is real warmup, not a synthetic loop.
Can I bring my own ICP list?
Yes. Upload a CSV or paste LinkedIn URLs. Mailer enriches and writes against your list rather than sourcing new ones.
Does it work in Hebrew?
Yes. Hebrew outbound is first-class — not a translation layer on top of English. The Hebrew reply rates in our sample are within 1-2 points of English.