We handed outbound to AI and reply rates fell. Where does AI actually help, and where does it backfire?
AI is very good at the research and routing underneath outbound and quite bad at being the voice of it. Use it to find accounts, watch for triggers like a funding round or a champion changing jobs, enrich your CRM and draft a first pass. Do not let it send. Buyers now recognise the texture of generated text and delete on sight, and campaigns run this way decay hard: one 2026 analysis puts reply-rate decay above 60 percent within eighteen months as recipients pattern-match the format. The tell is that the AI companies themselves are hiring more SDRs, not fewer, while the wider market cut SDR hiring 21 percent. Keep a human writing the first line and the last line.
Go deeper
4 resources, 4 link-checked.
📰 Newsletter
✓ Link checkedFreemiumAdvanced
Shows the specific places AI earns its keep in outbound (closed-lost re-engagement, champion job-change tracking, micro-campaigns off a signal cluster) rather than arguing about whether AI is good. The distinction between deterministic and agentic steps is the useful bit.
From
Growth Unhingedby Kyle Poyar and Brendan Short15 min read
Closed-lost re-engagement: an agent watches for signal clusters, pulls the last call transcript, finds the objection that killed the deal, then drafts the reopen.
Micro-campaigns: when signals stack on one account, build a 50 to 100 contact list with campaign-specific copy (example: 13 companies hiring Data Engineers who named a competitor).
Warm-path ranking: a relationship graph scores every possible intro into a target account by likelihood of success.
Tier 1 accounts route to a human for review; lower tiers send automatically.
The counterweight to every AI SDR pitch deck, with the failure modes named and numbered: sender reputation collapse, intent-data false positives at 31 to 47 percent, and the eighteen month reply-rate half-life. Read it before you sign an annual contract.
Actual job posting data instead of opinion: GTM posts down 15 percent, SDR posts down 21 percent, support down 37 percent, while GTM engineering headcount doubled and AI native companies more than doubled their SDR teams. It is the antidote to both the everything is automated and nothing has changed camps.
Recorded in 2026, once teams had lived through the reply rate drop. It draws the same line the answer does, AI for research, routing and a first draft, humans for the voice, and it comes from a company that sells outbound tooling, so it is not an anti-AI pitch.
