Sales is not the scarce resource
Studios can sell another brochure site. The constraint is downstream: discovery, design system exceptions, CMS schema, deploy, and the next six months of “small” content changes. Each project still behaves like a custom production even when the brief looks familiar.
Hiring does not fix that by itself. A larger team still copies last quarter’s stack, then spends the margin reconciling editor tools with the app. Retainers hide the cost until the calendar fills with work that cannot be productized.
Two different products, on purpose
On the CloudPack site we published a longer strategic reset: The 2026 agency fulfillment crisis. That essay is canonical for the engine argument. This post is MethodGate operating view.
When the deliverable should remain a marketing site, we do not want a new architecture every time. That is CloudPack — Next.js, CloudCannon, one schema for editors and the app. When the brief is a set of landings, that is CloudTemplates.
When the visitor should not hunt a nav at all, a template is the wrong instrument. That is Cintesa: conversational generative UI with a CMS, funnels, and agentic-web tools on the same project. Live at cintesa.run.
How we choose on a brief
- Editor-owned pages, campaigns, multilingual marketing → CloudPack.
- Fast landing variants with visual guardrails → CloudTemplates.
- Qualify, answer, and convert with conversational generative UI → Cintesa.
Forcing one of those into the other two jobs is how fulfillment gets worse, not better. Read the CloudPack essay for the engine thesis, then pick the product that matches the visitor’s job.
Also: why solo teams pick engines · the static site problem.
