Podcast — Inside AI-Powered B2B Outbound
Real talk on how AI agents find buyers, write personalised outreach, and book meetings without a human SDR.
Episode 1 — How NEO SDR Works
A deep dive into the NEO SDR agent system: how seven specialized AI agents replace an entire SDR team. We cover ICP fingerprinting, multi-signal prospect scoring, 5-step outreach sequences, reply classification, and the continuous learning loop that improves results over time.
Episode summary
In this episode we walk through the full NEO SDR pipeline from scratch. The system starts by building a precise Ideal Customer Profile (ICP) using semantic fingerprinting on the user's own website — no manual targeting required. That ICP drives Apollo search queries that surface hundreds of matching prospects, which are then scored on four dimensions: company fit, live intent signals, prior engagement, and email deliverability confidence.
We explain why deliverability is a first-class signal rather than an afterthought. Apollo's bulk-match API reveals verified business emails with deliverability guarantees for most B2B contacts. Prospects that cannot be verified are held back and periodically retried rather than sent and bounced. This keeps domain reputation high from day one.
The copywriter agent assembles fully personalised multi-step sequences — five emails per prospect — using company-specific pain points surfaced by real-time web research. Each email is unique: no templates, no merge-field spam. The outreachor agent dispatches them on a human-paced schedule, respects time zones, and stops the sequence the moment a positive reply arrives.
Reply classification runs automatically on every inbound message. The classifier distinguishes genuine interest, out-of-office, unsubscribes, wrong-person replies, and bounces. Positive replies trigger a calendar-link follow-up so the human rep only sees meetings ready to be taken.
Finally, the learner agent closes the feedback loop. Every reply — positive or negative — feeds a pattern library that the copywriter consults on the next campaign iteration. Over time the system learns which angles, subject lines, and CTAs convert best for each ICP segment.
Key takeaways
- Seven specialised AI agents (ICP Builder, Prospector, Researcher, Scorer, Copywriter, Outreachor, Learner) each own one job in the pipeline.
- Semantic ICP fingerprinting removes the need for manual prospect list building.
- Multi-signal scoring — fit × intent × engagement × deliverability — prioritises prospects most likely to reply.
- Email verification at the source (Apollo bulk-match) eliminates most hard bounces before a single email is sent.
- Every outreach sequence is unique per prospect; personalisation is driven by live web research, not generic templates.
- Automated reply classification routes positive intent to calendar booking with no human triage required.
- The learning loop makes each campaign measurably better than the last based on real reply data.
Timestamps
- 0:00 — Introduction: why AI outbound is different
- 2:15 — ICP fingerprinting with semantic analysis
- 7:40 — Apollo prospect sourcing and multi-signal scoring
- 14:20 — Email verification and deliverability gates
- 19:55 — AI copywriter: research, personalisation, sequences
- 26:10 — Outreach scheduling and send-time optimisation
- 31:45 — Reply classification and calendar booking
- 37:00 — The learning loop: how results improve over time
- 43:30 — Live results and what's coming next
What we cover across episodes
- How the NEO SDR agent system works end to end
- Multi-signal lead scoring and intent detection
- What makes outbound emails get replies in 2026
- Real campaign anatomy — ICP, sequences, reply handling
- Lessons from running live outbound for real customers
Watch the latest episode · How NEO SDR works · Apply for early access