Podcast Data And Workflow Spec

Full legacy body rendered for humans

Podcast Data And Workflow Spec

Runtime Data Objects

podcast_guests

Represents a person/company in the guest pipeline before or alongside an episode.

Recommended fields:

• id
• guest_name
• company
• title
• email
• linkedin_url
• status
• source
• notes
• invited_at
• booked_at
• bio_requested_at
• bio_received_at
• headshot_requested_at
• headshot_received_at
• last_followup_at
• created_at
• updated_at

podcast_episodes

Represents an actual episode from prep through publication.

Recommended fields:

• id
• episode_title
• guest_name
• guest_company
• topic
• status
• record_date
• publish_date
• riverside_link
• buzzsprout_episode_id
• buzzsprout_url
• youtube_video_id
• youtube_url
• transcript_url
• drive_folder_id
• show_notes_status
• content_kit_status
• approval_status
• notes
• created_at
• updated_at

podcast_handoffs

Tracks explicit internal handoffs.

Recommended fields:

• id
• episode_id
• from_agent
• to_agent_or_person
• handoff_type
• required_by
• status
• payload
• created_at
• completed_at

Workflow

guest_candidate
  -> guest_invited
  -> guest_booked
  -> intake
  -> pre_production
  -> ready_to_record
  -> recorded
  -> in_production
  -> ready_for_approval
  -> approved
  -> content_kit_ready
  -> content_kit_sent
  -> published

Stalled can be applied to any pre-published state with a reason.

Alert Rules

• intake: urgent until research and welcome/confirmation draft exist
• guest_booked without bio/headshot request: urgent after 1 business day
• bio/headshot request unanswered: follow-up due after 5 days
• ready_to_record with recording inside 3 days: watch
• ready_to_record with recording inside 1 day and missing logistics/assets: urgent
• recorded without transcript/production handoff: urgent
• ready_for_approval: urgent for Justin
• approved without content kit within 48 hours: urgent
• pre_production stale after 5+ days: watch; 10+ days urgent

Output Contracts

Daily digest must include:

• generated_at
• total_active
• count_by_status
• urgent_items[]
• watch_items[]
• pending_guests[]
• approvals_waiting[]
• handoffs_due[]

Each urgent item must include:

• entity type
• guest or episode name
• current status
• why it is flagged
• owner
• recommended next action

Open Questions

• Which Supabase project should own runtime podcast tables after the B2BEA.org rebuild: RedKey platform, B2BEA site database, or a dedicated B2BEA operations database?
• Should Jess create Gmail drafts directly, or should all external drafts go through a Studio/cockpit approval queue first?
• Does Mahesh still own production handoff, and what is his preferred folder/link structure?
• Should Buzzsprout and YouTube sync be part of Jess V1 rebuild or a later publishing slice?

Pre-Recording Readiness Fields

Add or derive these fields for scheduled episodes:

• recording_scheduled_at
• episode_folder_id
• headshots_folder_id
• prep_doc_id
• readiness_checked_at
• headshots_present
• prep_doc_reviewed
• reminder_sent_or_staged_at
• reminder_reason

One-Week Check Logic

A scheduled episode becomes eligible for readiness check when:

recording_scheduled_at <= now() + interval '7 days'

and

readiness_checked_at is null or readiness_checked_at < recording_scheduled_at - interval '7 days'

Headshots pass if Headshots and Photos Here contains at least one image file.

Prep doc review passes if there is a guest edit/comment signal available. If Drive revision/comment authorship is unavailable, Jess uses a weaker fallback: the doc exists and contains the generated context sections, then flags prep_doc_review_signal_unknown rather than claiming guest review.

Reminder is required when:

• headshots_present = false, or
• prep_doc_reviewed = false, or
• prep_doc_review_signal_unknown = true and Justin wants proactive reminders.

Post-Recording Data Model Addendum

Recovered full Jess design requires episode asset tracking after recording. Add or map these concepts:

episode_assets

One row per generated or delivered artifact. Required fields:

• id
• episode_id
• asset_type
• title
• status
• content or url
• source
• created_by
• created_at
• updated_at

Recovered asset types include: blog_justin, blog_guest, social_linkedin_justin, social_linkedin_guest, guest_content_kit, teleprompter_intro, youtube_metadata, buzzsprout_description, kb_entry, short_clip, thumbnail, quote_card, carousel, justin_prep_pack, teleprompter_questions.

episode_status_log

Every status transition should be logged before the episode row changes:

• episode_id
• from_status
• to_status
• changed_by = jess
• reason
• created_at

Post-recording statuses

• recorded
• in_production
• ready_for_approval
• published

Post-recording handlers must not skip status transitions.