Capability Lens

The Capability Lens examines why value flows or stalls by analyzing six layers of organizational capability. Each layer can either enable acceleration or create constraints that ripple across the entire value stream.

The Six Layers

Strategy & Culture

Vision, leadership commitment, and organizational culture around AI transformation.

People & Skills

Talent, capabilities, roles, and learning systems for AI delivery.

Ways of Working

Processes, methodologies, and collaboration patterns for AI initiatives.

Technical Platform

Infrastructure, tools, and technical capabilities for AI development.

Governance & Enablers

Policies, compliance, and organizational structures that enable safe AI adoption.

External Interfaces

Vendor relationships, ecosystem partnerships, and external data sources.

Strategy & Culture

Purpose: Align the organization around AI value creation with appropriate leadership commitment and cultural foundation.

Strategy & Culture sets the context for everything else. Without clear direction and cultural readiness, other capabilities cannot fully enable value delivery.

Five Dimensions

Dimension	Focus
Vision & Roadmap	Clarity of AI strategy and prioritized initiatives
Leadership Commitment	Executive sponsorship and resource allocation
Risk Appetite	Willingness to experiment and accept controlled failures
Learning Orientation	Commitment to continuous improvement and knowledge sharing
Change Readiness	Organizational adaptability and transformation capacity

How It Enables/Constrains

Enables when:

Clear AI strategy guides prioritization
Leadership provides resources and removes blockers
Culture encourages experimentation
Learning from failures is normalized

Constrains when:

No clear AI direction or competing priorities
Leadership disengaged or skeptical
Failure is punished rather than learned from
Change fatigue limits appetite for transformation

Flow Impact

Value Stream Phase	Strategy & Culture Impact
Discovery	Clear vision guides opportunity identification; learning culture enables experimentation
Delivery	Leadership support unblocks resources; risk appetite enables shipping
Validation	Learning orientation enables honest assessment; change readiness supports pivots
Foundations	Strategic commitment justifies platform investment

People & Skills

Purpose: Ensure the organization has the talent and capabilities required to deliver AI value.

AI initiatives require specialized skills alongside traditional engineering and product capabilities. This layer assesses whether the right people with the right skills are in place.

Five Dimensions

Dimension	Focus
AI/ML Expertise	Depth of machine learning and data science capabilities
Engineering Capability	Software engineering skills for production AI systems
Product & Design	Ability to translate AI capabilities into user value
Data Literacy	Organization-wide understanding of data and AI
Talent Development	Learning systems, career paths, and knowledge sharing

How It Enables/Constrains

Enables when:

Teams have the skills needed for their initiatives
Career paths attract and retain AI talent
Knowledge sharing scales individual expertise
Cross-functional collaboration is effective

Constrains when:

Critical skill gaps exist in key roles
Talent attrition outpaces hiring
Expertise is siloed in individuals
Cross-functional collaboration breaks down

Flow Impact

Value Stream Phase	People & Skills Impact
Discovery	Product/research skills drive opportunity identification
Delivery	Engineering and ML expertise determine delivery velocity
Validation	Data literacy enables meaningful impact measurement
Foundations	Platform engineering skills build shared capabilities

Ways of Working

Purpose: Establish effective processes and collaboration patterns for AI initiatives.

AI projects often require different approaches than traditional software development. This layer assesses whether processes enable or hinder AI delivery.

Five Dimensions

Dimension	Focus
Development Process	Methodology for AI/ML development (experimentation, iteration)
Collaboration	Cross-functional teamwork between ML, engineering, and product
Decision Making	How AI initiative decisions are made and escalated
Knowledge Management	Documentation, institutional memory, and knowledge transfer
Continuous Improvement	Retrospectives, feedback loops, and process evolution

How It Enables/Constrains

Enables when:

Processes support experimentation and iteration
Cross-functional teams collaborate effectively
Decisions are made at appropriate levels with appropriate speed
Knowledge is captured and shared systematically

Constrains when:

Rigid processes don’t accommodate AI’s iterative nature
Silos prevent effective collaboration
Decision-making is slow or centralized inappropriately
Knowledge walks out the door with individuals

Flow Impact

Value Stream Phase	Ways of Working Impact
Discovery	Collaboration processes enable cross-functional research
Delivery	Development processes determine velocity and quality
Validation	Feedback loops accelerate learning cycles
Foundations	Knowledge management ensures platform leverage

Technical Platform

Purpose: Provide the infrastructure and tooling required for AI development and deployment.

Technical capabilities determine what is possible and how efficiently teams can operate. This layer assesses the maturity of AI/ML infrastructure.

Five Dimensions

Dimension	Focus
Data Platform	Data availability, quality, cataloging, and access
ML Infrastructure	Training, experimentation, model management, and deployment
Development Environment	Tools, IDEs, notebooks, and development workflows
Production Systems	Monitoring, observability, and operational excellence
Security & Reliability	Security controls, reliability, and disaster recovery

How It Enables/Constrains

Enables when:

Self-service data access accelerates experimentation
ML infrastructure automates repetitive tasks
Development tools increase productivity
Production systems are reliable and observable

Constrains when:

Data is inaccessible or poor quality
ML infrastructure requires manual effort
Development tools are outdated or missing
Production issues consume team capacity

Flow Impact

Value Stream Phase	Technical Platform Impact
Discovery	Data access enables research; sandbox environments support prototyping
Delivery	ML infrastructure determines deployment velocity
Validation	Observability enables impact measurement
Foundations	Platform investment multiplies team effectiveness

Governance & Enablers

Purpose: Establish policies and structures that enable safe, responsible AI adoption.

Governance can either enable or constrain AI initiatives. Effective governance provides guardrails that allow teams to move quickly within defined boundaries.

Five Dimensions

Dimension	Focus
AI Ethics & Responsibility	Frameworks for responsible AI development and use
Compliance & Regulatory	Adherence to relevant regulations and standards
Data Governance	Data ownership, privacy, and usage policies
Risk Management	Identification, assessment, and mitigation of AI risks
Organizational Structure	Reporting lines, centers of excellence, and accountability

How It Enables/Constrains

Enables when:

Clear guardrails allow teams to move quickly within boundaries
Compliance is built into processes, not bolted on
Risk assessment is proportionate to actual risk
Organizational structure facilitates rather than hinders

Constrains when:

Governance is undefined, creating uncertainty
Compliance is an afterthought requiring rework
Risk aversion prevents reasonable experimentation
Organizational silos block collaboration

Flow Impact

Value Stream Phase	Governance & Enablers Impact
Discovery	Data access policies enable or block research
Delivery	Compliance requirements affect development and deployment
Validation	Ethics frameworks guide impact measurement
Foundations	Governance determines platform scope and boundaries

External Interfaces

Purpose: Leverage external partnerships, vendors, and data sources to accelerate AI value.

No organization builds everything internally. This layer assesses how effectively external resources are leveraged.

Five Dimensions

Dimension	Focus
Vendor Management	Relationships with AI/ML vendors and service providers
Partnerships	Strategic alliances for AI capabilities or data
External Data	Third-party data sources and market intelligence
Open Source	Utilization of open source tools and models
Ecosystem Engagement	Participation in AI communities and standards bodies

How It Enables/Constrains

Enables when:

Vendor relationships provide leverage without lock-in
Partnerships accelerate capabilities
External data enriches internal intelligence
Open source accelerates development

Constrains when:

Vendor lock-in limits flexibility
Partnership friction slows execution
External data is unavailable or low quality
Not invented here syndrome rejects external solutions

Flow Impact

Value Stream Phase	External Interfaces Impact
Discovery	Market intelligence informs opportunity identification
Delivery	Vendor tools and open source accelerate development
Validation	External benchmarks contextualize performance
Foundations	Build vs buy decisions shape platform strategy

Using the Capability Lens

Diagnosing Constraints

When a Value Stream phase is Blocked or Constrained, the Capability Lens helps identify root causes:

Identify the struggling phase (e.g., Delivery is Constrained)
Examine each capability layer for potential causes
Prioritize high-impact, addressable constraints
Design targeted interventions

Example Analysis

Situation: Delivery phase is Constrained - cycle times are 3x industry benchmarks.

Layer	Assessment	Constraint?
Strategy & Culture	Leadership supportive, culture open	No
People & Skills	ML engineers overloaded, hiring slow	Yes
Ways of Working	Code review bottlenecks, long approval chains	Yes
Technical Platform	CI/CD slow, deployment manual	Yes
Governance	Reasonable policies, clear guidelines	No
External	Good vendor relationships	No

Interventions: Address People (accelerate hiring, redistribute work), Ways of Working (streamline reviews), and Technical Platform (automate deployment).

Capability Interactions

Capabilities interact. Weakness in one layer often manifests as problems in another.

Common Patterns

Root Cause	Apparent Symptom
Strategy unclear	People frustrated, ways of working chaotic
Skills gaps	Platform underutilized, delivery slow
Governance restrictive	Platform limited, external interfaces avoided
Platform immature	Ways of working manual, delivery constrained

Next Steps

Value Stream - Understand what the Capability Lens diagnoses
Why Acceleration - The philosophy behind continuous improvement
Getting Started - Take the assessment