The Agentic Symphony

An interactive map of all the moving pieces in a production agentic AI system.

v2.1.0 · 38 nodes · 39 connections · 14 layers
🔗 View Data Model on GitHub

Connection Analysis

Autonomous

Agent-driven tool invocation — the agent intelligently chooses which tools to call and when.

Deterministic

Developer-wired tool invocation — tools are pre-configured at design time, not selected by the model.

Bidirectional

Two-way data flow — the connection carries data in both directions.

Transport

Data flows through a pipeline — post-processed and indexed by a separate system, not the agent.

The Agentic Loop

The most important insight hidden by linear diagrams: agent execution is iterative. After each tool result, the agent returns to the model for another reasoning step. The loop (reason → act → observe → reason) may execute many times per request. This loop is the defining characteristic that separates agentic from automated systems.

Autonomous vs. Deterministic Tool Invocation

The connection from the agent layer to tools is not uniform. Agents autonomously choose tools (non-deterministic). Pipelines and workflows use tools that are wired in at design time (deterministic). This is arguably the defining line between “agentic” and “automated” AI systems.

Memory is Not Monolithic

The architecture captures four distinct memory patterns:

Gateway Routing Pattern

In production, inference requests often route through a model gateway (OpenRouter, LiteLLM) before reaching the model endpoint. The gateway handles model selection, load balancing, failover, rate limiting, and cost tracking. A self-hosted LiteLLM instance can unify access to both cloud APIs and local model servers.

Safety Applied at Multiple Points

Safety guardrails don’t just constrain agents — they apply at three points: input filtering on prompts (injection detection), agent behavior constraints, and tool-level restrictions on MCP (preventing unauthorized actions).

HITL Varies by Agent Type

Human-in-the-loop patterns differ: per-action approval for agents (human approves each tool call), stage-gate for pipelines (review output before next stage), and exception-based for workflows (human notified only on low confidence).

Overlooked Value Routes

High-value data pathways that are simple but deliver outsized returns — typically invisible in architecture diagrams.

1. Prompts → Prompt Library

Good prompts are discovered through use, not designed upfront. Capturing and curating prompts that work well in production creates reusable institutional knowledge — “this is how we talk to models about X.”

User/System Prompts → Stored Prompts → Prompt Library

New agents and workflows can bootstrap from proven prompts instead of starting from scratch. It’s a form of organizational memory.

2. Conversation-Driven Context Enrichment

Conversations themselves are mined for context about the user — preferences, patterns, domain knowledge. A dedicated agentic workflow extracts, categorizes, and indexes this into the vector store.

User Prompt → Conversations (DB) → [Mining Workflow] → Vector Store → Context (RAG) → Future Prompts

This creates a second virtuous cycle distinct from generic context-mining:

The mining step is itself agentic — it uses LLMs to extract structured insights, not just raw embeddings.

3. Outputs → Knowledge Management

Every agent interaction potentially generates organizational knowledge. Without this route, that knowledge is trapped in individual conversation threads and lost. Agent outputs (research summaries, analyses, recommendations) flow to wikis and knowledge bases for organizational use.

Architecture Layers