Walk me through your architecture. Is this microservices? If it's a monolith, why should I believe it can serve enterprise customers?

The full executive answer

It is a deliberately chosen modular monolith running on serverless infrastructure — and I would defend that choice in front of any architect. One codebase, but internally split into 68 strictly separated domain modules, each with its own service layer, pure logic, and tests. Every one of our 414 API endpoints deploys as an independent serverless function on Vercel, so at runtime we actually get the scaling behaviour people want from microservices — each endpoint scales on its own — without the operational tax of running dozens of separately deployed services.

The industry has largely converged on this as the right pattern for our stage: Amazon Prime Video famously moved a service back from microservices to a monolith and cut costs 90%, and the 'monolith first' guidance from Martin Fowler says the same — you split into services when team boundaries and scale demand it, not before. We are a small team; microservices would multiply our failure modes, our network hops, and our on-call surface without adding a single unit of customer value today.

Crucially, the seams for future extraction already exist. Modules communicate through a typed internal event bus with a durable outbox — meaning every domain event is also written to a database table before it fans out — so if one module ever needs to become its own service, the boundary and the event contract are already drawn. That is 12-Factor discipline: stateless processes, config in the environment, backing services attached, one codebase deployed many times.