Studio
The design-time environment for enterprise intent, processes, capabilities, knowledge, decisions, and controls.
by Nu10PIE platform architecture
From intelligent intent to
enterprise action.
PIE makes controls and evidence structural: agents, decisions, human oversight, tools, and enterprise action operate as one system.
Illustrative architecture
Design intelligent work, execute it in context, and retain the evidence to examine it.
These five layers form the system’s functional frame. Governance and controls span each layer—not as a compliance box bolted beneath it.
05Governance & control planeIdentity · authorization · policies · risk tiers · approvals · lineage · evidence
01
StudioDesign-time environment for intent, process, capability, knowledge, decisions, and controls.
02
DeploymentControlled movement of system artifacts through environments.
03
RuntimeGoverned orchestration of the process as it executes.
04
Capability providersAgents, people, decisions, tools, APIs, data, and applications that perform work.
Evidence follows execution
Layer detail
Deployment
A controlled path for moving the configured system through the appropriate environments and release decisions.
Runtime
The governed orchestration layer that balances generative reasoning, deterministic logic, automation, and human judgment as the process requires.
Capability providers
The approved agents, human gates, decisions, tools, APIs, data services, and applications that perform work.
Governance & control plane
Identity, authorisation, policy, risk tiering, human approvals, lineage, observability, evidence, and operational visibility span the architecture.
Executable primitives, not the architecture
Capability providers do the work.
PIE governs how they are invoked.
An approved capability serves each process need, while enterprise requirements and control boundaries remain stable as implementations change.
AI agents
Reasoning and interpretation
Human-in-the-loop gates
Review and authorised approval
Decision nodes
Explicit rules and routing
MCP servers and tools
Capability invocation
API connectors
Enterprise-system interaction
Evidence mechanisms
Context and lineage capture
Capability as contract
“Assess counterparty risk” is a process need—not a permanent dependency on one implementation.
THE PROCESS ASKS FORAssess counterparty risk
AI agentInternal risk engineExternal APIHuman analystDeterministic rules
PIE routes each need to an approved provider, keeping governance, authorisation, and traceability intact as implementations change.
Executable architecture
Explore how PIE fits an enterprise system landscape.
See how PIE’s layers connect your capabilities into a governed enterprise operation with context and evidence.
Book a PIE Demo