No sequence to build
Skip the wizard step where you plan 7 emails and 3 LinkedIn messages. The agent reads the lead, the triggering signal and the conversation history, then writes the next touch at send-time, not in advance.
Fixed cadences treat every lead the same. Adaptive AI doesn't. For each lead, the engine reads the triggering signal, the conversation history and what converted on the agent's other leads, then picks the next channel, drafts the message and schedules the timing. Everything stays bounded by your touch ceiling and default delay, so the agent never runs forever.
Skip the wizard step where you plan 7 emails and 3 LinkedIn messages. The agent reads the lead, the triggering signal and the conversation history, then writes the next touch at send-time, not in advance.
You set a hard ceiling (default 7 touches per lead). Once hit, the lead moves to handoff toward your rep with full context. No infinite chases, no accidental spam.
You set a default delay between touches (48h by default). The AI aligns to it, can extend it when context calls for it (waiting for end of quarter, respecting an OOO), and is capped at 21 days max between two touches.
The planner picks the most relevant channel per touch (email first, LinkedIn for the follow-up, WhatsApp when you have the number, etc.). The router then applies your channel strategy (lock-on-reply, fallback) and per-channel caps, no collisions.
If the agent needs to know your offer (pricing tiers, FAQ link, value-prop, use cases), drop the pitch into the agent's "Product pitch" field. Injected as-is into the planner prompt, never rewritten.
On the agent's page, you see the full list of upcoming plans · channel, message preview, when, why. You inspect, you edit, you kill. Zero black box.
Sequence when you have a known winning cadence and want every lead to walk through it identically. Adaptive when leads come from heterogeneous signals (some are funding-round-fresh, some are job-changers, some are warm referrals) and one size doesn't fit all.
It reads the lead's full conversation history, the per-channel last-touch timestamps, the agent's default-delay-hours, and any business-hour gating. Then it proposes a nextActionAt timestamp, visible to you on the agent detail page before it fires.
Open the adaptive plan row, edit the next touch (channel, message, when), save. The planner respects your override and resumes from there on the next tick.
Yes. Adaptive vs Strict is a per-agent setting. Run an adaptive agent on warm inbound leads and a strict agent on a cold-email broadcast in parallel · they don't interfere.
Trial on request · contact our sales team.