A case study building a "reservation management + automated notifications" skill cassette with BALIA OS, for any business that takes bookings — restaurants, salons, physical therapy clinics, medical clinics, and more. We put together notifications to the business (via Discord) about incoming reservations, plus automated arrival reminders and thank-you messages to customers, customizing each piece through a business-specific interview.
Notifications to the business (Discord) and notifications to customers (WhatsApp, SMS, LINE, etc.) go through fundamentally different channels. In this case study, we sort out that distinction during the initial interview stage before moving on to design.
| Tool | Role | Think of it as |
|---|---|---|
| Chat Claude (claude.ai) | Decides "what to build" through conversation. Handles the design/spec side | A planning session with your architect |
| Claude Code in VS Code | Takes the finalized design and actually creates the files and writes the code | The contractor on site |
The first time you think through a reservation system, deciding on just a "business type" doesn't tell you what you actually need to manage. Based on what existing reservation systems (for restaurants, for salons) actually track, here's a starting point for the interview.
This is just a reference starting point. What you actually need varies by business and how your shop operates, so in Phase 0 we'll first ask about your business type, then use this as a base to work out the specific items to manage together.
In Phase 1, you'll decide between "building a new dedicated database + reservation site from scratch" or "continuing to use your current Google Calendar as the reservation ledger." As a reference point, here's a comparison using a hypothetical izakaya (20 seats, 4 tables) — what you can and can't do with each option.
Rule of thumb: If your reservation volume is low and you just want to try out notifications first, it's fine to start with calendar integration. If you want to properly use business-specific management items like visit history and allergy tracking, you'll need a dedicated database.
| Phase | Content | Tool Used | Deliverable |
|---|---|---|---|
| Phase 0 | Requirements interview (business type, reservation granularity, existing tools, notification channels) | Chat Claude | Requirements sheet |
| Phase 1 | Deciding the system architecture (new reservation site vs. existing calendar integration, choosing the customer notification channel) | Chat Claude | Architecture plan |
| Phase 2 | Building the cassette (SKILL.md/WORKFLOW.md and a dedicated SPEC/soul.md are all required) | Chat Claude | SKILL.md, WORKFLOW.md, SPEC/soul.md |
| Phase 3 | Implementation (iterative) | Claude Code in VS Code | A working cassette |
| Phase 4 | Testing and verification | Both | A verified cassette |
| Phase 5 | Polish and documentation | Chat Claude | Finished product + usage guide |
Since the assumptions vary a lot by business type, there's no ready-made TASK.md that skips the interview. Always start with the Phase 0 requirements interview.
Just paste this prompt as-is into a new conversation with "Chat Claude" (claude.ai) to kick off the Phase 0 requirements interview.
The rest is available in the full guide after purchasing T1.
(Full text including the detailed Phase 0 interview items, notification channel selection logic, and instruction-sheet templates for Phase 1 onward)
If you send automated messages to customers (arrival reminders, thank-you messages, etc.), make sure you obtain their consent at booking time — that "an automated message will arrive once the reservation is confirmed." We recommend designing the system to send only to reservations where consent has been confirmed.
The detailed Phase 0 interview items, notification channel selection logic, and instruction-sheet templates for Phase 1 onward — the full write-up is all laid out on the BALIA OS guide site.
See the Full Build Guide →Other case studies built the same way — community/fan site auto-updates, boat race prediction, crypto trading, and more — are live or in progress.
See All Case Studies →