What's actually changing in AI for people who do outreach — and why it matters for replies, deliverability, and pipeline. Original analysis, sources linked, no reposts. Last updated 2026-05-18.

2026-05-18 · deliverability · AI writing
LLM-written cold email is now easy to detect — and inboxes are starting toMass AI-generated outreach has a sameness signature. Filters and recipients both pattern-match it; the edge is shifting back to specificity, not volume.
2026-05-17 · AI SDR · automation
Autonomous 'AI SDR' agents: what actually changes vs. the hypeFully autonomous prospecting agents shift the bottleneck from sending to targeting and deliverability — not to 'set and forget' pipeline.
2026-05-16 · deliverability · filtering
Inbox providers are getting much better at scoring intent, not just spam wordsModern filtering scores whether mail behaves like wanted mail — engagement, sending patterns, authentication — making old spammy-keyword tricks irrelevant.
2026-05-15 · personalization · AI writing
AI personalization at scale: most of it is noise the recipient ignoresInserting a scraped detail isn't personalization. Recipients now discount obvious AI 'personalization'; relevance to a real problem is what still moves replies.
2026-05-14 · AI SDR · follow-up
AI reply triage moves the bottleneck to follow-through, not first contactWhen AI handles classifying and routing replies, the constraint becomes consistent, fast, human follow-through — the same gap that always loses deals.