12 SQLs per month for a niche AI speech company selling to contact centers
Tight, vertical ICP. Outbound built on specificity, not volume.
An 18-person AI speech startup had built a real-time agent assist product for contact centers. Their problem was reach: the buyers (VP Customer Experience, Director of Contact Center Ops) were on LinkedIn but unreachable through generic SDR outreach. We built a hyper-specific outbound motion that booked 12 qualified meetings per month at steady state.
THE CHALLENGE
What we walked into
- Crowded category — every prospect was being pitched by 5+ AI vendors per week.
- Buyers were senior (VP / Director level) and skeptical of cold AI pitches.
- Previous agency had generated 60+ 'meetings' that were 90% disqualified.
- Founder had no time to post or run outreach personally.
THE ICP
Exactly who we targeted
Job titles
Industries
Geographies
Company size
200+ contact center agents, $50M–$2B revenue
THE APPROACH
How we ran the motion
Pain-first ICP segmentation
We split the ICP into 3 pain buckets: 'AHT pressure', 'agent attrition', 'QA backlog'. Each got its own messaging track, opening hook and proof points. No prospect ever got a generic message.
Founder-led profile + ghostwritten posts
Founder's LinkedIn was rebuilt around one POV: 'Real-time AI for contact centers — without replacing agents'. We ghostwrote 3 posts/week in his voice. Engagement on his page 12x'd in 90 days.
Account-level multi-threading
Inside each target account we reached out to 2–3 personas (CX leader, Ops leader, IT/CCaaS owner). Sequencing ensured no two messages landed in the same week to avoid pattern detection.
Insight-led first message, not a pitch
Every first DM led with a contact center benchmark (e.g. 'AHT in mid-market health payers is now averaging 7m12s — saw {{company}} just expanded the {{location}} center, curious how you're handling it').
Tight qualification before booking
We added a 4-question async qualifier (size, current stack, timeline, budget owner) before a call was confirmed. Killed no-shows and brought disqual rate from 60% to 9%.
TRIGGER SIGNALS
What we waited for before reaching out
Every account had to fire at least one of these signals in the last 90 days before it entered the sequence. Trigger-based timing is the single biggest driver of reply rate.
New VP CX / Director Contact Center hire
LinkedIn job-change alerts
New leaders re-evaluate the AI stack in first 6 months
Contact center expansion / new site opening
Press releases + LinkedIn company posts
AHT, QA backlog and onboarding pain spike
Job openings for QA analysts / contact center agents (>10 roles)
LinkedIn Jobs scrape
Headcount strain — real-time assist becomes urgent
Migration to a new CCaaS (Five9, Genesys, NICE) in last 6 months
BuiltWith + LinkedIn skill tags
Stack is open — integrations easier to land
Engagement on competitor / category posts
LinkedIn engager scrape
Self-identified active buyers in market
THE FUNNEL
Numbers at each step of the motion
Monthly steady-state. Numbers rounded; conversion rates measured over 6+ months.
Target accounts (NA + UK)
200+ agent contact centers
900
Personas per account
CX + Ops + IT/CCaaS owner
2.6 avg
Connection requests sent / month
2 profiles · 100/wk each
800
Connections accepted
248 (31%)
Conversations started
152
Positive replies
29 (11.8%)
Async qualifier passed
18
Sales-qualified (SQL)
Disqual rate dropped 60% → 9%
12
CAMPAIGN MECHANICS
What the engine looked like
Profiles running outbound
2 (Founder + Head of Sales)
Connection requests / week / profile
100
Messaging tracks
3 pain-based variants
Personalization level
Account-level + persona-level
Tools
Sales Navigator, HeyReach, Clay, Apollo, Notion CRM
TIMELINE
From kickoff to steady state
Week 1–2
Founder positioning, pain segmentation, account list (900)
Week 3
Copy, sequences, qualifier built. Profile + content live.
Week 4–6
Ramp. First 14 SQLs booked.
Week 7–12
Steady 11–13 SQLs/month. 1st closed-won (mid-size BPO).
Week 13–24
Expansion into UK + Canada. 3 more closed-won.
Engagement length: Ongoing — 7 months at time of writing
THE RESULTS
Numbers that held month after month
SQLs booked
in 6 months
SQLs / month
months 3–6
Disqualification rate
down from 60%
Positive reply rate
of accepted connections
Closed-won
from first 68 SQLs
ACV influenced
first 6 months
"We'd been burned by an SDR agency before. What's different here is the qualification — every meeting that hits my calendar is a real buyer with a real budget. I don't waste an hour on a tire-kicker anymore."
Founder & CEO
AI Speech startup · 18 employees · Toronto
Want numbers like this for your pipeline?
Book a 30-minute call and we'll map an outbound motion built for your ICP.
Book A Call