Is your proposal related to a problem?
Cal.com captures attendee name, email, event type, and form responses at every booking but never activates that data after the confirmation email is sent. The host receives a calendar invite and nothing else. They have no context about who they are meeting, why the meeting was booked, or how to prepare. This means every host either manually Googles the attendee before the call or walks in unprepared.
No scheduling tool has solved this natively. Calendly has an AI Assistant but it requires manual prompting and only surfaces past meeting history — it does not generate a briefing from booking form data and deliver it to your inbox. Microsoft Copilot can prepare meeting briefs but pulls from emails and documents, not from the booking flow itself. The gap this PR fills — activating attendee data captured at booking to automatically prepare the host — is unclaimed by any scheduling tool natively.
Describe the solution you'd like
Add Booking Brief — an opt-in AI-generated briefing delivered to the host 30 minutes before each meeting. Using the Anthropic Claude API and the attendee data Cal.com already captures at booking time, the feature generates and emails the host:
- An attendee summary based on their name, email domain, and form responses
- Three suggested agenda items tailored to the event type
- A recommended opening question
- A tactical prep tip
The briefing is delivered via email through Cal.com's existing WorkflowReminder infrastructure and tasker queue. No new scheduler, no new database tables, no new infrastructure, no schema changes. The feature is completely invisible when ANTHROPIC_API_KEY is not set. Every existing booking flow is unaffected.
A working implementation is ready at: https://github.com/KevinNatera/Cal.com-Briefing-Engine/tree/kevins-branch
Describe alternatives you've considered
Manually Googling attendees before calls. Using external tools like Datagrid or Notion AI to generate briefings. Neither is integrated into the booking flow and both require manual effort from the host on every call, defeating the purpose.
Additional context
From a business perspective, this feature gives Cal.com a concrete reason to gate a meaningful AI capability behind paid tiers. At the Organizations plan ($37/user/month), the inference cost to run Booking Brief for a 10-person team at 1,000 monthly bookings is under $8/month — less than 2% of plan revenue. At the Teams plan ($15/user/month), the same team at 1,000 bookings costs under $8/month in inference. It also strengthens the enterprise argument: HIPAA-regulated industries would want briefing data to stay on-premise, which only the self-hosted enterprise tier enables. That is a direct upsell lever that requires no additional engineering beyond what this implementation already delivers.
The implementation is 367 lines across 8 files with four passing Vitest unit tests and updated developer docs. Inference cost per briefing using the Anthropic API with prompt caching is under $0.004, negligible at every tier.
Requirement/Document
Working branch with full implementation, tests, and docs: https://github.com/KevinNatera/Cal.com-Briefing-Engine/tree/kevins-branch
Is your proposal related to a problem?
Cal.com captures attendee name, email, event type, and form responses at every booking but never activates that data after the confirmation email is sent. The host receives a calendar invite and nothing else. They have no context about who they are meeting, why the meeting was booked, or how to prepare. This means every host either manually Googles the attendee before the call or walks in unprepared.
No scheduling tool has solved this natively. Calendly has an AI Assistant but it requires manual prompting and only surfaces past meeting history — it does not generate a briefing from booking form data and deliver it to your inbox. Microsoft Copilot can prepare meeting briefs but pulls from emails and documents, not from the booking flow itself. The gap this PR fills — activating attendee data captured at booking to automatically prepare the host — is unclaimed by any scheduling tool natively.
Describe the solution you'd like
Add Booking Brief — an opt-in AI-generated briefing delivered to the host 30 minutes before each meeting. Using the Anthropic Claude API and the attendee data Cal.com already captures at booking time, the feature generates and emails the host:
The briefing is delivered via email through Cal.com's existing WorkflowReminder infrastructure and tasker queue. No new scheduler, no new database tables, no new infrastructure, no schema changes. The feature is completely invisible when ANTHROPIC_API_KEY is not set. Every existing booking flow is unaffected.
A working implementation is ready at: https://github.com/KevinNatera/Cal.com-Briefing-Engine/tree/kevins-branch
Describe alternatives you've considered
Manually Googling attendees before calls. Using external tools like Datagrid or Notion AI to generate briefings. Neither is integrated into the booking flow and both require manual effort from the host on every call, defeating the purpose.
Additional context
From a business perspective, this feature gives Cal.com a concrete reason to gate a meaningful AI capability behind paid tiers. At the Organizations plan ($37/user/month), the inference cost to run Booking Brief for a 10-person team at 1,000 monthly bookings is under $8/month — less than 2% of plan revenue. At the Teams plan ($15/user/month), the same team at 1,000 bookings costs under $8/month in inference. It also strengthens the enterprise argument: HIPAA-regulated industries would want briefing data to stay on-premise, which only the self-hosted enterprise tier enables. That is a direct upsell lever that requires no additional engineering beyond what this implementation already delivers.
The implementation is 367 lines across 8 files with four passing Vitest unit tests and updated developer docs. Inference cost per briefing using the Anthropic API with prompt caching is under $0.004, negligible at every tier.
Requirement/Document
Working branch with full implementation, tests, and docs: https://github.com/KevinNatera/Cal.com-Briefing-Engine/tree/kevins-branch