Show Booking

Show Booking

Overview

Execute an end-to-end workflow for showing requests:

1. Parse intake from free-form prompt or email text.
2. Build per-listing call jobs.
3. Hand off call execution to the tour-booking sub-agent.
4. Generate calendar invite files from confirmed slots.
5. Return a concise confirmation summary.

Inputs

Collect these fields before running outbound calls:

• Client full name.
• Listings (address, listing ID if present, office phone, listing office/agent name if present).
• Preferred windows and timezone.
• Booking constraints (lockbox notes, occupants, minimum notice).
• Confirmation target (email/SMS destination for status updates).

If any listing is missing a phone number, flag it as blocked and do not place calls for that listing.

Workflow

1) Parse intake

Run:

python3 scripts/intake_request.py --input-file /path/to/intake.txt --output /tmp/showing-intake.json

Or pass inline text:

python3 scripts/intake_request.py --input-text "Book showings for ..." --output /tmp/showing-intake.json

2) Build call queue

Run:

python3 scripts/orchestrate_showings.py --intake /tmp/showing-intake.json --output /tmp/showing-plan.json

This produces:

• call_queue: listings with phone numbers ready for calls.
• blocked: listings missing required data.
• calendar_candidates: records ready for invite creation after call confirmation.

3) Delegate calling to tour-booking

For each call_queue record, invoke tour-booking/scripts/place_outbound_call.py with:

• Listing metadata.
• Preferred windows.
• Client identity.
• Callback instructions.

If live calling is not approved, run with --dry-run and return the generated payload.

4) Create invites for confirmed slots

When a listing returns a confirmed date/time:

python3 scripts/create_invite_ics.py \
  --input /tmp/confirmed-showings.json \
  --output-dir /tmp/showing-invites

The script emits one .ics file per confirmed showing. Import into Google Calendar or send directly as attachments.

5) Return status

Report:

• Confirmed showings with time, address, and invite file path.
• Pending callbacks.
• Blocked listings and the missing field(s).
• Total calls attempted and success/failure counts.

Guardrails

• Explicitly identify the caller as an AI assistant acting for the brokerage/realtor.
• Respect local telemarketing and consent requirements.
• Keep a full audit trail: request payload, call result, booking outcome, and timestamps.
• Never claim a showing is confirmed until the call result explicitly indicates confirmation.