Key Concepts
The mental model for how agenté. works.
C-Suite Seats
Each seat is a specialized AI agent with deep domain expertise. They're not generic chatbots — they have defined roles, guardrails, and capabilities.
The Routing Model
You talk to one interface — your COO. It understands your intent and routes to the right specialist automatically. Multiple seats can be engaged simultaneously for cross-functional questions.
Think of it like a real executive team: you brief the COO, and they coordinate the rest.
Synthesis vs. Drill-Down
There are two modes of interaction:
- Synthesis — "Give me the big picture on X." The COO pulls insights from multiple seats and presents a unified view.
- Drill-down — "Show me the numbers on that." You dig into a specific seat's domain for depth.
Most conversations flow between both modes naturally.
Memory & Context
Your C-Suite has persistent memory across sessions. It remembers:
- Your business context, entities, and structure
- Past decisions and their outcomes
- Your communication style and preferences
- Ongoing projects, deadlines, and commitments
Memory is organized in layers — from quick-reference facts to deep session history — so the right context is always available without overload.
Context Files
Each workspace has a set of context files that the C-Suite reads at the start of every session:
- SOUL — Personality, tone, behavioral rules
- MEMORY — Running operational memory
- USER — Your preferences and communication style
- AGENTS — Fleet routing (who handles what)
These files ensure every session starts with full context — no cold starts.
Approval Gates
The C-Suite never ships without your approval. This is structural, not optional:
- Documents go to your drafts folder. You review and send.
- Decisions are presented as options + recommendation. You choose.
- External communications are drafted, never sent directly.
- Financial actions require explicit confirmation.