Increasing AI Maturity: Navigating the AI Stages of Adoption with the Five Pillars

In my previous article on the AI Stages of Adoption (AISA), I outlined how organisations progress through their AI journey—from Experimenting to Adopting, Optimising, Transforming, and ultimately Scaling. Since publishing that piece, many readers have asked the same follow‐up question: “How do we know when we’re truly ready to move from one stage to the next?”

It’s a crucial question. I’ve seen organisations celebrate early pilot successes and immediately attempt to leap from Experimenting straight to Optimising, only to encounter significant challenges. The reality is that AI adoption isn’t merely about completing pilots or deploying new technologies—it’s about developing fundamental organisational capabilities that enable sustainable progress.

By measuring your maturity across five capability domains that I call the Five Pillars, you can accurately assess your readiness to transition between AI stages of adoption and identify critical gaps before they derail your journey.

Why Capability Pillars Matter

AISA provides a valuable macro view of organisational AI maturity, illustrating the progression from initial experimentation to ecosystem-wide scaling. However, the stages themselves don’t specify the underlying capabilities required for successful transitions.

Through the early work I did on the AWS Cloud Adoption Framework, I knew that explaining progression through capability domains was the right approach. For AISA, I’ve identified five capability domains that determine an organisation’s readiness to advance through the AI adoption journey:

1. Governance & Accountability
2. Technical Infrastructure
3. Operational Excellence
4. Value Realisation & Lifecycle Management
5. People, Culture, & Adoption

These pillars aren’t separate stages but capability domains that cut across every level of AI maturity. I regularly encounter organisations with considerable strength in some areas (such as technical foundations) but significant weaknesses in others (like governance or culture). The tension between these strengths and weaknesses often determines whether an organisation’s transition to the next adoption stage will be smooth or painful.

Navigating Stage Transitions

In my experience advising boards and leadership teams, I’ve observed that transitions between AI stages rarely follow a neat checklist. Instead, organisations successfully advance when certain foundational capabilities align. Let me share what I’ve seen across those transitions:

1. Experimenting to Adopting - This transition typically involves moving from ad hoc pilots to a more formalised approach with dedicated AI budgets, initial governance structures, and cross-functional teams. Without sufficient maturity in the Governance & Accountability pillar, organisations often find themselves with proliferating “shadow AI” initiatives that create significant risk exposure. Likewise, inadequate Technical Infrastructure can limit an organisation’s ability to move beyond isolated experiments.

2. Adopting to Optimising - The leap to Optimising requires organisations to standardise AI efforts, embed them into core processes, and measure ROI more rigorously. This transition demands strength in both Technical Infrastructure and Operational Excellence pillars, along with robust Value Realisation capabilities. Often otherwise promising AI programs stall when organisations attempt this jump without the operational discipline to monitor value outcomes or the processes to consistently measure business impact.

3. Optimising to Transforming - This critical transition is where AI evolves from an operational tool to a driver of business model innovation. Success depends heavily on the People & Culture pillar, as transformation requires widespread AI literacy and cultural acceptance. Additionally, the Value Realisation & Lifecycle Management pillar becomes crucial for identifying transformative opportunities. Without these foundations, even organisations with strong technical capabilities struggle to achieve meaningful business transformation.

4. Transforming to Scaling - The final transition extends AI beyond organisational boundaries to broader ecosystems. This requires excellence across all pillars, with particular emphasis on Governance & Accountability mechanisms that can span partner networks and value chains, and Technical Infrastructure that can support ecosystem-wide integration. At this stage, weakness in any pillar can create systemic vulnerabilities.

The Five Foundational Pillars

Let’s examine each pillar in detail to understand its role in enabling successful stage transitions.

Pillar 1: Governance & Accountability

This domain ensures your AI initiatives operate within safe and legal boundaries while maintaining clear human oversight and responsibility. It includes managing AI-specific vulnerabilities, handling privacy and regulatory obligations, preventing misuse or harmful outputs, establishing protocols for crisis response, and ensuring AI decisions can be explained and traced to accountable parties. This combined pillar incorporates frameworks for human-in-the-loop processes, escalation thresholds, and transparency mechanisms like model cards.

When organisations move from Experimenting to Adopting, governance and accountability requirements increase exponentially. The informal oversight and basic policies that sufficed for limited pilots become inadequate when multiple business units begin deploying AI. Organisations rushing this transition will face regulatory scrutiny or reputational damage from poorly governed adoption with unclear lines of accountability. As organisations approach the Transforming stage, the stakes rise further. When AI reshapes core business processes, weak governance and accountability mechanisms can undermine the entire transformation effort. Without clear ownership of AI decisions, organisations struggle to maintain trust with stakeholders as AI becomes more deeply embedded in critical business functions.

Pillar 2: Technical Infrastructure

This domain addresses the foundational technology elements that support AI systems. It focuses on data architecture, compute resources, and the underlying platforms that enable AI development and deployment. This includes data storage and processing capabilities, cloud or on-premises infrastructure, networking, security controls, and the technology stack that supports AI workloads.

As organisations move from Experimenting to Adopting, the technical infrastructure requirements shift from supporting isolated experiments to enabling enterprise-grade AI systems. The makeshift environments suitable for proofs of concept become insufficient when supporting multiple AI initiatives across the organisation. When progressing toward the Optimising stage, robust technical infrastructure becomes critical for consistent performance and scalability. Without properly architected data storage, efficient compute resources, and secure development environments, organisations find themselves unable to operationalise AI at scale.

Pillar 3: Operational Excellence

This pillar focuses on the processes, practices, and methodologies for effectively running AI systems in production. It covers data quality and governance, continuous model monitoring, Machine Learning Operations (MLOps) practices, sustainable AI operations, and the day-to-day management of AI systems. This domain ensures that AI solutions function reliably, consistently, and efficiently once deployed.

The journey from Adopting to Optimising represents a significant shift from proving AI’s value to operationalising it at scale. This transition demands mature MLOps practices, real-time monitoring for model drift, effective data quality procedures, and standardised operational processes. Without excellence in these areas, organisations find themselves handling each AI initiative as a separate project rather than building cumulative operational capability. When progressing to the Transforming stage, organisations cannot reliably underpin large-scale business transformation with inconsistently operated AI systems. Without robust operational practices, even technically sound AI solutions will fail to deliver sustained value at scale.

Pillar 4: Value Realisation & Lifecycle Management

This domain ensures organisations capture measurable business value from AI investments. It spans the full lifecycle—from identifying high-value use cases to managing vendor relationships, protecting intellectual property, and measuring ROI.

The shift from Experimenting to Adopting typically hinges on demonstrated business value. If organisations cannot show tangible gains from initial pilots, it becomes nearly impossible to secure the expanded budgets and executive support needed for broader adoption. When organisations attempt to move from Optimising to Transforming, the value realisation requirements intensify. AI must now generate new revenue streams or business models, not just incremental improvements. This transition requires sophisticated processes for identifying, prioritising, and scaling the most promising AI initiatives.

Pillar 5: People, Culture, & Adoption

This pillar addresses the human dimension of AI adoption—change management, skills development, and cultural readiness. It’s about fostering an AI-literate organisation where teams see AI as an enhancement rather than a threat.

The transition from Adopting to Optimising requires broadening AI skills beyond specialist teams. During early adoption, a small group of data scientists might suffice, but optimisation demands that business teams understand how to effectively work with AI solutions. By the time organisations approach the Transforming and Scaling stages, they need a deep bench of AI expertise at all levels—from technical specialists to business leaders and frontline employees. Without investing in change management and skills development, organisations find that their sophisticated AI systems go unused or are actively circumvented.

Implementing the Framework

In practical terms, organisations should develop specific maturity indicators for each pillar. For example, under Governance, an “Initial” level might mean having basic ethical guidelines, while an “Advanced” level would include comprehensive risk assessment frameworks, monitoring protocols, and regular governance reviews. By periodically assessing against these criteria, you’ll identify which pillars need strengthening before attempting the next stage transition. This creates a practical roadmap for prioritising investments and capability building.

When your assessment highlights weaknesses, focus investments accordingly—perhaps enhancing data governance capabilities, launching an AI literacy program, or creating standardised risk management frameworks. By strengthening the appropriate pillars, you remove roadblocks that would otherwise impede your progress.

The Role of an AI Centre of Excellence

In the early days of enterprise cloud adoption at AWS, I developed our initial thought-leadership on establishing the Cloud Centre of Excellence (CoE). AI requires the same set of oversight and guardrails to orchestrate pillar maturity development. An effective CoE serves as both guardian of responsible AI practices and accelerator of AI innovation, coordinating across governance, accountability mechanisms, technical standards, and cultural engagement. In my next article, I’ll detail how to structure an AI Centre of Excellence that balances these responsibilities effectively. The right CoE structure enables organisations to progress through the AI Stages of Adoption while maintaining appropriate controls and maximising value creation.

Building Bridges for AI Advancement

The AI Stages of Adoption framework helps organisations understand their position on the journey from initial experimentation to ecosystem-wide scaling. But how successfully you navigate each transition depends on the strength of your five capability pillars.

Neglecting these foundational elements might allow temporary progress but typically leads to painful setbacks when governance failures, technical gaps, operational issues, or cultural resistance emerge. By systematically building capability across all five pillars, organisations create solid foundations for sustainable advancement through each stage of AI maturity.

Think of each pillar as a critical support structure for the bridge between one stage and the next. The stronger these supports, the more confidently your organisation can cross into new territory without fear of collapse. By measuring and strengthening these capabilities, you create a clear, actionable roadmap for AI adoption that balances ambition with pragmatism.

In my next article, I’ll explore how an AI Centre of Excellence can orchestrate this capability development, ensuring your organisation advances through the AI adoption stages with confidence and control.

About the Author

Mario Thomas is a transformational business leader with nearly three decades of experience driving operational excellence and revenue growth across global enterprises. As Head of Global Training and Press Spokesperson at Amazon Web Services (AWS), he leads worldwide enablement delivery and operations for one of technology's largest sales forces during a pivotal era of AI innovation. A Chartered Director and Fellow of the Institute of Directors, and an alumnus of the London School of Economics, Mario partners with Boards and C-suite leaders to deliver measurable business outcomes through strategic transformation. His frameworks and methodologies have generated over two-billion dollars in enterprise value through the effective adoption of AI, data, and cloud technologies.