Stage 1 — AI-Curious: Awareness without action. Stage 2 — AI-Experimenting: Isolated pilots, no structure. Stage 3 — AI-Adopting: Selective deployment in specific functions. Stage 4 — AI-Scaling: Structured, organisation-wide rollout. Stage 5 — AI-Optimising: Continuous improvement of AI-enabled processes. Stage 6 — AI-Embedded: AI is core to strategy, operations and competitive advantage.
Most organisations currently sit at Stage 2 or early Stage 3.
Why maturity models matter — and how to use this one
A maturity model is only useful if it helps you make decisions. Too many frameworks describe stages in abstract terms that are difficult to map to operational reality. This one is built differently: each stage is defined by observable signals — things you can actually look for in your organisation — and by the specific blockers that typically prevent progress to the next stage.
The six stages are not a linear prescription. Organisations often have different maturity levels in different functions: a finance team at Stage 4 while operations remains at Stage 1. The goal is not to advance every part of the organisation at the same pace, but to understand where the highest-value opportunities are and what is genuinely holding progress back.
Before walking through each stage, it's worth naming the five dimensions we assess across all of them. Maturity isn't a single score — it's a profile.
Strategy & Leadership
Is AI part of the business strategy? Are senior leaders actively sponsoring and visibly using AI?
Data & Infrastructure
Is data accessible, clean, and governed? Does the technical infrastructure support AI deployment?
People & Skills
Do staff have the confidence and capability to use AI effectively in their roles?
Governance & Risk
Are there policies, oversight structures, and risk management practices that allow AI to be deployed safely?
Process & Operations
Have workflows been redesigned to incorporate AI, or is AI being added onto existing processes?
Culture & Adoption
Does the organisation's culture support experimentation, learning, and change?
The six stages
AI-Curious
The organisation is aware of AI — leaders are reading about it, attending events, and fielding questions from the board. There may be a handful of individuals using AI tools personally, but there is no organisational strategy, no structured approach, and no investment beyond the exploratory conversations happening at leadership level.
This stage is characterised by energy without direction. The interest is genuine, but it hasn't yet translated into a decision about what to do, where to start, or who is responsible. AI is a topic of conversation, not a programme of work.
- AI appears on board agendas as a discussion item but not as a strategic priority
- Individual employees using ChatGPT or Copilot without organisational awareness
- No AI policy, governance framework, or formal assessment of AI opportunities
- Technology decisions are being deferred pending "more clarity" on the AI landscape
- Uncertainty about where to start — AI feels too broad
- Risk aversion at leadership level without a clear framework for managing AI risk
- No internal owner with the mandate and capability to drive an AI programme
AI-Experimenting
The organisation is running pilots. There are one or more teams actively testing AI tools — often in isolation from each other, often without a shared framework for evaluation, and typically without a clear path from pilot to production. The pilots generate enthusiasm and some genuine insight, but they don't connect to a broader strategy.
This stage is characterised by activity without coherence. Multiple experiments may be running simultaneously without coordination. Early results are positive but not systematically captured. The organisation is learning, but the learning isn't accumulating into organisational capability.
- Two or more separate AI pilots running in different parts of the business
- No shared evaluation criteria across pilots — each team defines success differently
- Positive pilot results but no defined process for scaling to production
- AI tools being evaluated by IT and/or individual teams without a cross-functional view
- Growing pressure from the board or CEO to "show progress" on AI
- No AI strategy to connect pilots to business priorities
- Data quality and access issues that surface only once pilots are underway
- Insufficient change management investment — technology is advancing faster than people capability
AI-Adopting
The organisation has moved beyond experiments. Specific AI tools are in active use in one or more functions, with genuine productivity or quality improvements to point to. There is visible leadership support, a growing community of capable users, and early signs of cultural shift — people are talking about AI as part of how they work, not just as a project.
This stage is characterised by selective success. The organisation knows AI works in some contexts. The challenge is that success is concentrated in particular functions or teams, and it hasn't yet generalised. There are capable early adopters and resistant non-adopters, and the gap between them is widening.
- One or two functions with proven, measurable AI deployment
- Internal case studies and success stories beginning to circulate
- A named AI lead or working group with some formal responsibility
- An AI usage policy in draft or in place
- Training has been delivered but confidence levels vary significantly across teams
- Inconsistent manager commitment — some teams thrive, others barely engage
- Lack of a scaling playbook — success in one function hasn't been documented in a replicable way
- Governance gaps — policies and oversight haven't kept pace with deployment
AI-Scaling
AI is being deployed at scale across the organisation, with a structured programme, defined ownership, and consistent governance. The organisation has moved from asking "should we do this?" to "how do we do this well and fast?" Adoption is broad enough that AI capability is beginning to differentiate the organisation's performance.
This stage is characterised by structured momentum. The organisation has a playbook, a governance framework, and a growing base of capable users. The challenges are now operational: sustaining pace, managing the inevitable resistors, and ensuring that quality and risk standards hold as deployment accelerates.
- A formal AI adoption programme with executive ownership and a dedicated team
- AI deployed across multiple functions with measurable outcome data
- A functioning AI governance framework including risk management and oversight
- Regular AI-related communication from senior leadership — not just programme updates
- Internal AI champions network operating across the organisation
- Fatigue — large-scale adoption programmes are demanding and momentum can stall
- Data infrastructure limitations that constrain more sophisticated AI use cases
- Talent gaps in AI-adjacent skills as ambitions move beyond off-the-shelf tools
AI-Optimising
AI is embedded in core processes and the organisation is now focused on continuous improvement: refining how it uses AI, building more sophisticated applications, and integrating AI outputs into decision-making at senior levels. The question has shifted from "how do we get people to use AI?" to "how do we get more value from our AI capability?"
This stage is characterised by compound returns. Each improvement in AI capability builds on the last. The organisation is developing proprietary workflows, fine-tuning models on its own data, and beginning to see AI as a source of competitive advantage rather than a productivity tool.
- AI outputs regularly informing senior-level decisions
- Custom AI workflows built on top of foundation models, tailored to organisational context
- Dedicated AI engineering or data science resource building proprietary capability
- AI investment decisions made on ROI data, not on competitive pressure
- External recognition of AI capability as a differentiator
AI-Embedded
AI is core to how the organisation works, decides, and competes. It is not a programme or a function — it is part of the operating model. New hires are expected to be AI-capable. Strategic decisions are informed by AI analysis. Products and services have been redesigned around AI capability. The organisation would be materially less effective without it.
This stage is characterised by strategic integration. AI is no longer something the organisation does — it is part of what the organisation is. Very few organisations have reached this stage. Those that have tend to be technology-native or have made sustained, multi-year investments in AI capability building.
- AI capability is referenced in the organisation's public strategy and investor communications
- Job descriptions at all levels reference AI proficiency as a core requirement
- Products or services that would not exist without AI capability
- The organisation is attracting talent specifically because of its AI capability
- AI governance is embedded in the board's risk and audit remit
How to move through the stages faster
The most common mistake organisations make is trying to leap stages. A Stage 2 organisation that attempts a Stage 4 rollout without the governance, change management, and data infrastructure to support it typically creates expensive chaos and then retreats — often back to Stage 1.
The second most common mistake is treating stage progression as a technology problem. In our experience, the bottleneck at every stage up to Stage 5 is people, culture, and governance — not the technology. The tools are available. The question is whether the organisation can absorb and use them.
Three principles that accelerate maturity regardless of stage
- Fix the data before the tool. More AI programmes stall on data quality and access than on any other single factor. Understand your data landscape before committing to a technology.
- Move managers first. Frontline managers determine the cultural reality of adoption. An organisation of willing managers progresses faster than an organisation of capable individual contributors who can't get buy-in from their line manager.
- Measure outcomes, not activity. Stage progression is measured by what AI does for the business, not by how many tools are deployed or how many licences are activated. Define your outcomes before you start and measure them obsessively.
Where most organisations are right now
Based on our work across sectors and organisation sizes, the distribution in mid-2025 looks roughly like this: the majority of organisations sit at Stage 2 or early Stage 3. A meaningful minority — particularly in financial services and technology sectors — are at Stage 3 to 4. Stage 5 and above remains rare outside technology-native organisations.
The practical implication: most organisations are in the stage where the critical decisions are about how to move from pilot to production. The technology choices matter less at this point than the adoption strategy, the governance framework, and the change management investment.
If you're uncertain where your organisation sits, the five-dimension assessment above is a starting point. Better still, take our free AI Readiness Assessment — it gives you a scored view across all five dimensions and tells you specifically where your biggest gaps are.
Find out where your organisation sits.
Our free AI Readiness Assessment scores you across strategy, people, data, governance, and culture — and tells you specifically what's holding you back.
Take the free assessment →