The Periodic Cube of AI: Making Sense of the AI Landscape
The Periodic Cube of AI is a framework for understanding, assessing, and navigating the complex world of artificial intelligence.
Table of Contents
The Challenge: AI is Overwhelming
If you’ve tried to understand AI beyond the headlines, you’ve probably felt lost.
Foundation models, MLOps, embeddings, RAG, guardrails, feature stores… the terminology alone can be paralyzing.
And that’s before you even consider how these pieces fit together, who should own them, or whether to build, buy, or integrate them.
For executives, the challenge is strategic:
- Where should we invest?
- What capabilities do we need?
- Are we focusing on the right things?
For practitioners, it’s tactical:
- What should I learn next?
- How does my expertise connect to the bigger picture?
I created the Periodic Cube of AI to answer both questions with a single, interactive 3D visualization.
Why a Cube? (And Why "Periodic"?)
The original periodic table revolutionized chemistry by organizing elements along multiple meaningful dimensions. Suddenly, patterns emerged. Relationships became visible. Predictions became possible.
I applied the same principle to AI—but I needed more dimensions. A flat table couldn’t capture the richness of the AI landscape. So I built a cube.
The Periodic Cube organizes 60 production-related AI components (and 60 corresponding learning components) across eight visual dimensions:
Spatial Position (X, Y, Z)
X-Axis: Functional Groups The horizontal axis organizes components by their role in the AI stack:
- Data & Infrastructure – The foundation: data sources, compute, storage
- Model Development – Training, serving, and managing AI models
- Tooling & Integration – Developer tools, APIs, testing frameworks
- Application & Orchestration – Agents, workflows, user interfaces
- Governance & Operations (MLOps) – Security, compliance, monitoring
Y-Axis: SFIA Categories The vertical axis aligns with the Skills Framework for the Information Age (SFIA):
- Strategy and Governance
- People and Change
- Process and Operations
- Technology
- Data
Z-Axis: Technology Readiness Level (TRL) The depth axis shows maturity:
- Emerging – Cutting edge, high risk, high reward
- Maturing – Gaining adoption, patterns stabilizing
- Established – Battle-tested, well-understood
Visual Properties (Color, Size)
Color: Build, Buy, or Integrate Each component’s color indicates the typical sourcing strategy:
- Orange (Build) – Custom development required
- Blue (Buy) – Commercial solutions available
- Green (Integrate) – Open source or API integration
- Purple (Hybrid) – Mix of approaches
Box Dimensions: Three More Variables
- Width → Criticality (Mission-Critical to Enhancing)
- Height → Human Intensity (Fully Automated to Human-Driven)
- Depth → Organizational Ownership (who typically owns it)
Production Components: The 60 Building Blocks
The cube’s production layer represents the components you’d find in a mature enterprise AI platform. Here are some examples:
Data & Infrastructure
- Data (Systems of Record, 3rd-Party Providers) – Your raw material
- Compute (GPU/Silicon, Cloud, Edge) – The engine
- Enterprise DWH/Data Lake – The warehouse
- Data Catalog & Lineage – The map
Model Development
- Foundation Models – The brain (often bought or accessed via API)
- Training Data Store – The curriculum
- Feature Store – Pre-computed intelligence
- Model Hub & Inventory – Your model portfolio
Tooling & Integration
- Developer Tools (SDKs, IDEs) – The workbench
- Context Catalog – Dynamic knowledge injection
- Test Harness – Quality gates
Application & Orchestration
- Agent Framework – Autonomous reasoning systems
- RAG Pipeline – Grounded, context-aware responses
- Workflow Orchestration – Multi-step AI processes
- Conversational Interfaces – The user-facing layer
Governance & Operations
- Guardrails & Safety – Behavioral boundaries
- Model Registry – Version control for models
- Compliance Framework – Regulatory alignment
- Audit & Forensics – Accountability trails
Learning Components: The Knowledge Map
Behind every production component lies knowledge to acquire. The cube’s learning layer mirrors the production layer, but adds three new dimensions:
Domain
- Perception & Data
- Intelligence & Reasoning
- Action & Control
- Ethics & Society
Scale
- Nano – Fundamentals (transistors, signals, basic concepts)
- Micro – Building blocks (features, embeddings, neural networks)
- Macro – Systems (agents, applications, governance)
Depth
- Theory – Conceptual understanding
- Build – Hands-on implementation
- Product – Operational excellence
For example, if you’re working with Foundation Models (Production), the related learning components include:
- Neural Weights (Concepts) – Theory level
- Model Training (Implementation) – Build level
- Embeddings (Ops) – Product level
This creates a personalized learning path based on where you work in the stack.
Six Perspectives, One Cube
Different stakeholders care about different dimensions. I designed the cube to support six persona views:
CEO View
«Strategic alignment & business value»
- Focus: Mission-critical components, competitive advantages
- Key insight: Which AI investments differentiate us?
CTO View
«Technical architecture & scalability»
- Focus: Engineering ownership, integration complexity, emerging tech
- Key insight: Where do we need deep technical capability?
CFO View
«Cost optimization & ROI»
- Focus: Build vs. buy decisions, cost structures
- Key insight: Where are we over-investing? Under-investing?
CPO View
«Product velocity & innovation»
- Focus: Product-facing components, time-to-market
- Key insight: What’s blocking faster AI feature delivery?
CISO View
«Security & compliance»
- Focus: Mission-critical security, data governance
- Key insight: Where are our AI risk exposures?
COO View
«Operational efficiency»
- Focus: Automation opportunities, human-intensive processes
- Key insight: Which processes are ripe for AI-driven transformation?
Self-Assessment: Your AI Maturity Mapped
- Rate your current capability for each relevant component (1-5)
- Set target levels based on your strategic goals
- Visualize gaps directly on the cube
- Outer cube = Target maturity
- Inner cube = Current capability
- Color coding shows gap severity (green = on track, amber = close, red = needs attention)
Why This Matters Now
AI is no longer optional. But success requires more than enthusiasm—it requires clarity:
- Clarity on components: What are the building blocks?
- Clarity on relationships: How do pieces connect?
- Clarity on ownership: Who is responsible for what?
- Clarity on maturity: What’s proven vs. experimental?
- Clarity on gaps: Where do we need to invest?
The Periodic Cube of AI provides all five.
Try It Yourself
Explore the interactive 3D Periodic Cube of AI at: ai-periodic-cube.webmemo.ch
Take the AI self-assessment here.
Related to the Periodic Cube of AI
- Learning paths by persona: How a marketing specialist’s journey differs from a product manager’s
- The research behind the cube: Academic frameworks and industry standards that informed the design
