Bug Reports
Open

Ability for AI to follow up after conversation dries out

Migrated from Canny.

Canny Post ID: 697b9a9b73f1123c40caf6a5

Canny URL: https://spur.canny.io/admin/board/feature-requests/p/ability-for-ai-to-follow-up-after-conversation-dries-out

Category: Improvement

Original status: open

Original votes: 3

Original comments: 0

Original author: Royal Tomar

Created at: 2026-01-29T17:36:27.861Z

Original request details:

Currently, customers have to work around this by creating an AI Action that triggers on every message to schedule follow-ups (see: Follow_Up_Bot workflow pattern). This approach has several limitations:

Inefficient - The action fires on every single message, even when the conversation is actively ongoing

No true "conversation went cold" detection - It's just a timed delay from the last message, not intelligent detection of disengagement

Complex setup - Requires building a full workflow with delays, working hours checks, and conditional logic

What we need:

Native AI capability to detect when a conversation has gone inactive/cold (configurable inactivity threshold, e.g., 20 mins, 1 hour, 24 hours)

AI autonomously sends a contextually relevant follow-up without requiring a trigger on every message

Respects working hours automatically

Different follow-up strategies based on conversation stage (e.g., prospect vs. existing customer)

Use case: Health/wellness businesses like Earlyfit want to nurture leads who went quiet after initial interest, without spamming active conversations.

1 Comment

Sign in to comment

firstTeamยทabout 16 hours ago

REQUEST_SUMMARY_V2 Problem - Currently, customers have to work around this by creating an AI Action that triggers on every message to schedule follow-ups (see: Follow_Up_Bot workflow pattern). This approach has several limitations: Inefficient - The action fires on every single message, even when the conversation is actively ongoing No true "conversation went cold" detection - It's just a timed delay from the last message, not intelligent detection of disengagement Complex setup - Requires building a full workflow with delays, working hours checks, and conditional logic What we need: Native AI capability to detect when a conversation has gone inactive/cold (configurable inactivity threshold, e.g., 20 mins, 1 hour, Expected behavior - Allow this condition/action directly in workflow builder - Support testing with real sample events before publishing - Provide clear logs for why a workflow did or did not run Business impact - Reduces manual operational effort for teams - Improves consistency and customer experience - Helps teams move faster with fewer support escalations Legacy category: Improvement

Sign in to comment