On/Off Toggle for AI Follow Up

As the purpose of AI Follow Up goes beyond the purpose of call recording/transcription, where customers object we need to offer CSMs a granular way to turn AI Follow Up off, while keeping recording and transcription on, on a call-by-call basis.

Currently, there is no specific option for AI follow-up at the user  or even at the admin end, apart from the set up. We do not fully understand the use case where a call is recorded and transcripts are allowed to be generated but should not be logged into Timeline with Summary?
To build something with granular user permissions for summarizing calls selectively at this point is a very heavy lift and doesn’t seem ideal. 
The only option to not have it summarized is to not have it recorded and transcribed. 

Introducing something where despite a call is recorded and transcribed but should not be summarized by AI Follow-up, is going to be sizeable effort and is not on our near term roadmap.

Hi ​@ChiragJ,

Thanks for the updated.

The core use case is that recording/transcription and AI Follow-Up fall under different processing purposes, and our customers treat them differently from a privacy and consent perspective.

• Recording and transcription are typically approved because they support basic business operations: accurate notes, context retention, and internal record‑keeping. Most customers already expect this.

• AI Follow-Up, however, performs additional automated processing—summaries, insights, enrichment, extraction, and writing back derived data into systems. This is considered a separate processing purpose under privacy frameworks like GDPR. Many customers explicitly object to AI-based analysis even when they allow recording/transcription.

Because these purposes must be independently consented to, our CSMs need the ability to disable only the AI processing on a per-call basis when a customer raises an objection—without losing the benefits of recording/transcription, which remain allowed.

Without a per‑call toggle, we are forced to either:

1. violate the customer’s preference, or
2. disable recording/transcription entirely (which harms operational needs)

This is the gap we’re trying to address.