How we used AI to generate $2M in pipeline from a single email sequence

One sequence, one segment, one outcome

We didn't run a complicated multi-channel campaign. We picked one underserved segment — B2B SaaS companies between 50 and 200 employees that had recently raised a Series A — and built a single five-email sequence targeted directly at their pain. The AI wrote it. We shipped it. $2M in pipeline in 11 weeks.

Here's exactly how we did it.

Step 1 — Defining the segment

Series A SaaS companies are in a specific moment: they've just raised, they're under pressure to grow fast, and they're usually understaffed relative to their ambitions. That's a wedge. We used Crunchbase + Clay to pull 4,200 companies that matched the profile, enriched them with tech stack data, and filtered down to 1,800 that were likely to benefit from our product.

Step 2 — The AI writing the sequence

We prompted the model with three inputs: the segment profile, three anonymised customer success stories from similar companies, and our core value proposition. The model generated five email variants per slot in the sequence. We A/B tested two per slot and let it run.

The winning subject lines from email 1:

• Winner: "How [Company] could cut their sales cycle in half"

• Variant B: "A question about your post-Series A growth stack"

• Variant C: "Saw you raised — here's what your fastest peers are doing differently"

Step 3 — Send strategy

We didn't blast the full list at once. The agent sent in cohorts of 200 per day, monitored open and reply rates in real time, and paused sending if deliverability signals degraded. Sending windows were Tuesday–Thursday, 7–9am recipient local time.

sendCohort({
  size: 200,
  window: { days: ['tue', 'wed', 'thu'], hours: [7, 9] },
  timezone: 'recipient_local',
  pauseIf: { bounceRate: 0.03, spamRate: 0.001 }
})

sendCohort({
  size: 200,
  window: { days: ['tue', 'wed', 'thu'], hours: [7, 9] },
  timezone: 'recipient_local',
  pauseIf: { bounceRate: 0.03, spamRate: 0.001 }
})

sendCohort({
  size: 200,
  window: { days: ['tue', 'wed', 'thu'], hours: [7, 9] },
  timezone: 'recipient_local',
  pauseIf: { bounceRate: 0.03, spamRate: 0.001 }
})

Sequence structure

"Email 5 — the breakup — generated 22% of all replies. Don't skip it. People respond when they think it's the last time you'll reach out."

Results

• 1,800 contacts in sequence

• 41% open rate on email 1

• 6.2% reply rate across the full sequence

• 112 discovery calls booked

• $2.1M in pipeline created in 11 weeks

• $18 cost per pipeline dollar generated

The biggest lesson: specificity beats volume. 1,800 highly targeted contacts outperformed a previous blast to 22,000 generic leads by every metric.