# Current State

## Core operating systems in use today

### Launch27
Used as a major booking and scheduling system.
It remains important operationally, but has major limitations.

Known realities:
- useful for static and reference data
- not reliable enough through the official API for live booking and schedule operations
- live availability and booking changes often require direct UI use, Playwright automation, Airtable support, or manual verification
- there are existing scripts in the workspace for Launch27 investigation and rescheduling

### GoHighLevel (GHL)
Used mainly as a communications and CRM hub.
Current view is that GHL is archaic and not ideal as the long-term operating core.
It may remain useful as a communications layer in the near term, but is not an inspiring long-term system of record.

### Airtable
Used for important operational data support.
Important caution:
- the Airtable "active" flag for customers is not reliable
- upcoming confirmed or pending bookings are a better operational truth source for customer activity status

### Twilio + Vapi + Leah
Leah is the AI receptionist initiative.
Current role:
- SMS flow is live through Twilio Serverless
- voice staging is live via Vapi
- production voice is not yet fully rolled out

The strategic direction for Leah is highly relevant to CleanOS because it represents the AI interface layer for customer interaction.

### Other current systems and concerns
- QuickBooks for authoritative month-end accounting
- YNAB for more real-time financial visibility
- Slack for team notifications and ops coordination
- Google tools and service-account data access for reporting and operations
- HubSpot is actively being considered as a possible CRM option

## Operational pain and fragmentation

The current business uses multiple overlapping tools.
This creates several strategic problems:
- fragmented customer and booking data
- unclear system of record boundaries
- manual work and duplicate entry
- brittle automation between tools
- poor fit between generic tools and cleaning-specific workflows
- difficulty making AI agents act reliably because the operational truth is split across systems

## Current-state framing for CleanOS

The opportunity is not simply to build a prettier app.
The opportunity is to decide:
- what should remain in the stack
- what should be integrated
n- what should be replaced
- what system should own the canonical operational truth

## Questions that still need investigation

- What exactly does Launch27 still do best today?
- Which Launch27 workflows are mission-critical and hardest to replace?
- Could HubSpot become the CRM layer while another system owns booking and scheduling?
- Which current tools are worth keeping because replacement cost is too high right now?
- Where does data actually originate for leads, contacts, quotes, bookings, and follow-up?