high-pressure-performance-first
We build with the understanding that restaurants are a high-pressure, time-is-gold industry. Every architectural decision prioritises fast response times: aggressive caching at the query and view layer, background Celery tasks for anything that does not need to block the request cycle, WebSocket-driven real-time updates so UIs never need to poll, and lean serializers that avoid N+1 queries. Staff and kitchen crews cannot wait for slow screens during a lunch rush — the platform must be as fast as the kitchen.
transparent-errors-to-customers
We are transparent with error messages and surface them to the customer together with the objects being affected. Customers can track organization-level errors and understand the state of their data or orders when something goes wrong. This aligns with our debug-first architecture: fail hard, surface the problem clearly, and give everyone — developers and customers alike — the context they need to understand and resolve issues quickly.
no-full-ai-customer-automation
We do not fully automate customer conversations with AI. AI cannot replace the human touch required for quality customer service in the restaurant context. Instead, AI is used to assist staff — rewriting messages, detecting intent, suggesting replies — while a human remains in control of every outgoing conversation. This keeps interactions warm, accurate, and on-brand without sacrificing the speed benefits of AI tooling.