How it works
From cold prospect to booked meeting, on autopilot.
What it looks like end to end. Real numbers from a B2B SaaS client: 4,454 connections, 778 replies.
Stage 1 — Targeting
We build the list of who to reach.
ICP fingerprint to Sales Nav filters to enrichment to exclusions. Refreshed every two weeks as the campaign learns.
Your ICP
ICP fingerprint
We start with who, not what.
Onboarding nails industry, title, company size, geography, and the buying signals that mean a prospect is ready. The B2B SaaS client landed on VPs and Directors at 50-500 person SaaS companies, US/CA only.
Sales Nav filters
Translate the fingerprint into a real query.
Industry, headcount, function, seniority, geography, plus anti-targets. We tune the query until pulled records match the ICP at over 90%.
Enrichment
Add what LinkedIn doesn't show.
Company funding stage, recent hiring spikes, tech stack, growth signals. These become personalization hooks later, not just filters.
Exclusions
Remove competitors, customers, do-not-contact.
Suppression lists merged on first send and refreshed weekly. We never message someone twice across the same campaign.
Continuous refresh
The list isn't static.
Bi-weekly we replace exhausted segments with fresh records. If a sub-segment outperforms, we double down. If one underperforms, we cut it.
Stage 2 — Outreach
Connection request, then a 3-touch DM sequence.
Personalized opener, then three messages spaced across two weeks. Sent at human-pace from warmed profiles.
Daniel Reyes
Director of Sales · B2B SaaS
Connection accepted
52% acceptance · industry avg 25%
Day 0 · You
Liked your post on rep ramp-up — the 60-day ramp number you cited mirrors what we see across our SaaS book. Open to swap notes?
Day 3 · You
Quick one — when your AEs follow up after a demo, are they writing the email themselves or working off a template?
Day 7 · You
Connection request
Short, specific, no pitch.
One line that signals you read their profile. Acceptance rates lift from 25% to 50%+ when the opener feels human.
DM #1 — Day 0
Value-led, no ask.
Sent the day they accept. Frames the problem we solve in their language. Zero pressure to book.
DM #2 — Day 3
Qualifying question.
Asks one specific thing about their stack or process. Forces a yes or no. Replies here are the highest-quality leads.
DM #3 — Day 7
Soft CTA, then break up.
Offers the call. If no reply, we close the loop politely. We don't keep poking past three messages.
Send schedule
Human-pace, not blast-pace.
5-7 sends per profile per day, randomized intervals, business hours only. Profiles stay warm and don't trip LinkedIn's spam detection.
Stage 4 — AI Auto-ReplyThe differentiator
The AI writes back. Sounds like you. Books the call.
Reads the full conversation, picks the right tone, drafts the reply, sends after a short human-like delay. Hands off when it should.
Inbound reply
Sarah Chen
VP Marketing · Altair Data
“Interested. How does this compare to what we're running with SalesReach?”
Replied 2 min ago
AI draft
Calendar
Sarah Chen ↔ Lumos
Tue · 2:00 PM PT
Full conversation context
Reads the whole thread, not just the latest line.
Knows what was said, what tone you've been using, and what the prospect's last objection was. No reset between turns.
Tone matching
Warm for interested, deflecting for objections.
An interested prospect gets a confident yes-and. An objection gets a calm reframe. The AI picks based on the classification.
Real-time draft
Types out a reply in seconds.
You can watch it think. Each draft is logged so we can audit and improve over time.
Humanized send delay
Not instant. Not a bot.
Replies fire 15 to 90 minutes after the prospect's message, jittered. Looks like you stepped away from your desk and came back.
Knows when to hand off
If a prospect drops a calendar link or asks for pricing in detail, the AI stops.
We pull a human in. The hot moments don't get auto-replied to.
Self-improving
Winners get scored. Copy evolves.
Replies that produce a booked call get tagged. The AI biases toward what's working for your specific ICP, not a generic average.
Stage 5 — Booked Meeting
The call lands on your calendar. We tell you it's coming.
Direct calendar handoff, hot-lead Slack alert, CRM sync, weekly digest. You show up to the call. We do everything else.
Calendar
Daniel Reyes ↔ Lumos
Tue · 2:00 PM PT
Slack alert
Hot lead pinged
CRM sync
Logged in HubSpot
Weekly digest
Sent every Friday
Calendar handoff
Cal.com, Calendly, HubSpot Meetings, ChiliPiper.
If you have a public booking link, we route to it. If you don't, we propose times and confirm both ways.
Hot-lead alert
Slack ping the moment it's confirmed.
Prospect name, title, company, the conversation thread. You walk into the call already up to speed.
CRM sync
Logged in HubSpot, Attio, or wherever you live.
Conversation history, classification, source campaign. No manual data entry.
Weekly digest
Every Friday, the numbers.
Connections sent, replies, classifications, meetings booked. With commentary on what's working and what's about to change.
The result
4,454 connections. 778 replies. 36.5% reply rate.
All from a system the founder didn't have to manage. The numbers above are from a real B2B SaaS client we ran this for.
200+ businesses scaled
FAQ
Common questions.
Ready to scale outbound?
30-minute call. We'll audit your current outbound and map out a custom campaign plan.