Knowledge and memory
Separate durable knowledge from session, workflow, episodic, and organizational memory.
Enterprise AI architecture · First edition
Enterprise Intelligence Architecture
A reference architecture for governed, memory-centric enterprise AI systems.
This book examines the system responsibilities that appear when AI moves from experimentation into operational use. It treats intelligence as an architecture problem spanning knowledge, memory, reasoning, orchestration, governance, execution, and validation.
Version 1.0 manuscript and architecture are frozen. Paperback proof approval is in progress.
01
Book overview
Enterprise Intelligence Architecture is a technical reference for designing AI systems that must work inside real organizations.
The book moves beyond model selection and prompt design to the responsibilities that make an enterprise system dependable: preserving context, routing work, enforcing policy, recording decisions, handling exceptions, and validating outcomes.
The problem it addresses
A capable model does not provide memory, authority, accountability, or operational control. These concerns belong to the surrounding architecture. The book provides a shared structure for reasoning about them before they become production failures.
02
Intended audience
Written for people responsible for moving AI systems beyond demonstrations.
- Software and platform architects Defining system boundaries, shared capabilities, and integration patterns.
- Technical leaders Evaluating architecture decisions, delivery risk, and operating responsibility.
- Platform and software engineers Building the services around models, data, workflows, and enterprise controls.
- Governance and operations teams Establishing policy, review, observability, recovery, and auditability.
03
What readers will learn
The book develops a practical vocabulary for the responsibilities of an enterprise intelligence system.
Reasoning and routing
Match work to capabilities, constraints, and control paths instead of treating routing as a cost decision alone.
Orchestration and execution
Design workflows that preserve state, handle exceptions, and complete work across enterprise systems.
Governance and authority
Define policy boundaries, approval responsibilities, and accountable human control.
Validation and trust
Use evidence, traceability, review, and outcome verification to support defensible decisions.
Architecture synthesis
Connect the layers into a reference architecture that can be adapted to existing platforms and operating models.
Key topics
- Enterprise intelligence architecture
- Memory-centric systems
- Model routing
- Policy and governance
- Authority models
- Evidence and decision artifacts
- Observability and validation
- Production failure patterns
04
Companion resources
Source material and implementation work associated with the architecture.
Reference implementation
Open Intelligence Platform
The Open Intelligence Platform is the primary implementation reference for concepts introduced in the book, including model routing, enterprise memory, knowledge collections, gateway services, workspace management, and architecture decisions.
Purchase links
Companion repositories
Release information
- Target
- Version 1.0.0
- Manuscript
- Frozen
- Digital edition
- Leanpub
- Paperback
- Proof approval in progress
Related books
- AI-Native Software Development Lifecycle
- Agentic Engineering
- Context Engineering