Does a call answered by AI Receptionist generate a RingSense (ACE) domains/pbx/insights event?

We are integrating RingCentral voice into an internal application and want to receive call transcripts programmatically rather than downloading them from the Admin Portal.

We can see the ACE / RingSense Insights event filters and intend to subscribe to:

/ai/ringsense/v1/public/accounts/~/domains/pbx/insights

using the RingSense app scope and the ReadCompanyCallRecording permission.

What we cannot determine from the documentation is whether calls handled by the AI Receptionist are covered.

Specifically:

1. When a call is answered by an AI Receptionist assistant, rather than by a human user on a RingEX seat, does that call generate a domains/pbx/insights event?

2. If it does not by default, is there a configuration that makes it do so, such as enabling Automatic Call Recording on the extension the AI Receptionist is connected to? The ReadCompanyCallRecording permission requirement suggests that insights are derived from call recordings, so we want to confirm whether an AI-answered call is recorded in a way that RingSense can analyse.

3. If AI Receptionist calls are out of scope for RingSense entirely, is there any other supported way to receive AIR call transcripts at an external endpoint? We have found only Admin Portal viewing and per-call .TXT downloads. The AIR Integrations gallery also appears to offer only CRM and calendar connectors, with no custom webhook option.

4. If AI Receptionist calls do produce insights, which identifier should be used to correlate them? We ask because there is an unanswered thread from seven months ago, "RingSense Get Recording Insights returns 404," where a licensed, correctly scoped app passed a valid recording ID to domains/pbx/records/{id}/insights and received RAH-3001 "Insights not found for the provided recording", despite the transcript being visible in RingSense. We would rather not run into the same issue. If the webhook delivers the correlation ID directly in its payload, that would answer this question.

For context, we are not trying to transcribe our own media. We know the Audio & Video AI asynchronous speech-to-text API accepts a ?webhook= callback for media we submit. We specifically want RingCentral's own transcript for a call handled by its AI Receptionist.

Thanks.