Skip to content

L16 - Application & Domain Logic

L16 covers the product, workflow, domain rules, user experience, business process, and application-specific value created on top of the lower layers. It is the part users usually experience directly.

L16Application & Domain Logic
  1. Workflow
  2. Domain rules
  3. User experience
  4. Outcome ownership

What belongs here

L16 is where AI systems become products, copilots, agents, research workflows, internal tools, creative systems, or domain-specific applications. It is not less technical than lower layers; it simply asks different questions about usefulness, accountability, and fit.

Representative products and domains

Product or domain Why it might fit Adjacent layers
ChatGPT General AI assistant product built on model, tool, memory, and governance layers. L8 context, L14 memory, L15 safety
Claude Assistant product and API surface with model, tool, and safety behavior visible to users. L10 tools, L15 safety
Microsoft Copilot Product family embedding AI into productivity workflows. L14 identity, L16 workflows
GitHub Copilot Developer-assistance product with IDE, codebase, and workflow integration. L9 retrieval, L10 tools
Cursor AI coding environment where application UX and domain workflow are central. L8 context, L10 tools
Perplexity Answer and research product where retrieval, citations, and UX combine. L9 retrieval, L16 product
Harvey Domain-focused AI product for legal and professional services workflows. L15 governance, L16 domain logic
Dollhouse Collection Catalog and activation product where discovery, install flows, and artifact handling define the user experience. L9 retrieval, L11 registry
Elemental Surveys Applied research workflow where domain framing, review paths, and outcome ownership sit squarely in the application layer. L8 context, L15 governance
Merview Companion review and rendering utility that makes outputs, drafts, and artifact inspection usable to people. L15 governance, L16 workflow
DollhouseMCP User-facing runtime surface where many lower-layer capabilities become a coherent product and workflow. L10 tools, L14 memory

Boundary questions

  • When does an agent framework remain L12 orchestration, and when does it become an L16 application?
  • How should domain expertise, evaluation, and human review be represented without hiding lower-layer dependencies?
  • Are "copilots" a product pattern, a domain pattern, or a bundle of many lower layers?

Signals to watch

  • Domain-specific AI products developing their own evaluation and governance expectations.
  • Application UX making lower-layer decisions visible through citations, approvals, memory controls, and tool traces.
  • Organizations asking for stack-level transparency behind AI product behavior.