The 5 Stages of AI Maturity: Where Does Your Business Sit?

AI use has become routine in business. According to McKinsey’s 2025 State of AI survey, 88% of organisations use AI in at least one business function, yet nearly two-thirds have not started scaling it across the enterprise.

That gap between use and business impact is where AI maturity makes the biggest difference.

AI maturity looks at how effectively AI is integrated into workflows, decision-making, reporting and governance. A business can have widespread AI adoption and still be relatively immature if AI remains limited to isolated tasks or individual teams.

Different AI maturity models use different criteria, but they broadly track the same progression: from early experimentation to AI becoming part of how the business operates.

For revenue teams, we recommend approaching AI maturity through five stages: 

1. Awareness
2. Experimentation
3. Operational Adoption
4. Integrated Transformation
5. AI-Powered Operations

This framework is designed to help businesses plan what needs to be in place as AI moves from individual use into established revenue operations.

Why AI maturity matters

McKinsey’s 2025 global survey covered organisations across industries, regions and company sizes. Only around 6% of respondents qualified as AI high performers.

Those organisations showed some clear operational differences:

• They were nearly three times more likely to redesign workflows when introducing AI.
• They were three times more likely to have strong senior leadership ownership of AI initiatives.

The research points to something important: getting more value from AI requires changes to the way work is managed, not simply wider access to the tools.

For example, an AI summary may save a salesperson a few minutes before a call. The wider value depends on whether the CRM contains reliable information, the right lead reaches the right salesperson, and the outcome improves future scoring or reporting.

AI maturity looks at the full process surrounding the task.

Stage 1: Awareness

At the Awareness stage, AI use is led by individual interest.

Employees experiment with public tools, share prompts, and find small ways to reduce repetitive work. Usage varies significantly. Some employees use AI every day, while others remain uncertain about what they may use or which information they can safely provide.

The business may know AI use is happening without having a clear view of the tools involved or the data being shared.

Stage 1 signifiers:

• Employees using personal AI accounts
• Departments testing tools independently
• AI use focused mainly on drafting and summarising
• Limited guidance on data privacy
• No agreed way to assess value

The priority at this stage is visibility. The business needs to understand which tools employees already use and where AI produces useful results.

From there, leaders should establish basic AI governance before experimentation expands. That includes practical usage guidelines, clarity on what company or customer data may be shared with AI tools, and someone responsible for overseeing AI adoption.

Stage 2: Experimentation

At the Experimentation stage, teams begin running formal AI pilots.

Marketing may test AI-assisted campaign production. Sales might introduce call summaries or prospect research. Service teams may explore an AI agent, while RevOps tests data enrichment or workflow automation.

But what comes after the pilot?

A tool may complete the task successfully, but the business has not decided how that task should fit into normal work. Results are often based on employee feedback rather than agreed performance measures. Adoption remains inconsistent.

Progress requires better use-case selection. Teams should begin with a defined operational problem and understand how the current process works before introducing AI.

They should also check whether the data supporting the use case is reliable enough to scale, who can access it, and what level of human review will be required. These decisions are easier to make during a controlled pilot than after AI has already been added to a live workflow.

A useful pilot should answer a few practical questions:

• Which process are we improving?
• What result should change?
• Who owns the outcome?
• What data will the AI use, and is it reliable enough?
• What would need to happen for this to become part of normal work?

Without that structure, businesses can run many pilots without building much AI maturity.

Stage 3: Operational Adoption

Stage 3 begins when AI becomes part of a repeatable business process, rather than something individual employees use when they remember to.

In HubSpot, this could include using Breeze to:

• Summarise CRM records before sales calls
• Categorise records inside workflows
• Support lead prioritisation using agreed criteria

By this stage, the data and governance work started earlier becomes more important because AI is now influencing repeatable processes. Incomplete or inconsistent CRM data can weaken summaries, recommendations and automated actions, which is why HubSpot recommends standardising properties, managing duplicates and consistently capturing customer interactions.

MIT CISR found that the biggest financial shift occurs when organisations move from AI pilots to scaled ways of working.

For many businesses, moving from Stage 2 to Stage 3 will deliver more value than introducing another AI tool.

Stage 4: Integrated Transformation

At Stage 4, AI-supported processes begin working across marketing, sales, and customer service.

In HubSpot, this could include:

• Using engagement data to prioritise and route leads
• Connecting AI-supported workflows across the customer journey
• Using forecasting data as part of regular revenue planning

Governance now needs to extend across teams and connected workflows. Teams need clear rules around which data informs recommendations, where human review is required and how AI-supported processes are measured.

HubSpot’s AI forecasting feature is a good example. A more mature organisation uses the projection as part of a defined forecasting process, compares it with actual performance and improves the deal data affecting its accuracy.

Stage 5: AI-Powered Operations

At Stage 5, AI forms part of how the business manages revenue operations.

Teams use current data to identify risks, recommend actions, and adjust processes while work is taking place.

An AI-powered RevOps model may identify pipeline risk before a formal review, update lead prioritisation as new information becomes available, or use customer service activity to inform account-management decisions.

People remain responsible for decisions involving customer relationships and commercial risk. AI provides analysis, context, and suitable automation, while employees apply judgment where the situation requires it.

The organisation also continues to monitor its AI-supported processes. Prompts need review, workflows change, and governance must keep pace with new use cases.

Data quality also remains an ongoing responsibility. As AI becomes more deeply embedded, businesses need to keep reviewing the information feeding those systems rather than treating data preparation as a once-off implementation task.

At this stage, the competitive advantage comes from the organisation’s ability to use its own data and processes to make better decisions more quickly. Competitors may buy the same software, but they cannot easily reproduce the operating model behind it.

AI adoption and AI transformation

AI adoption measures whether people are using AI tools.

AI transformation considers whether that use has changed how the business operates.

A company can have hundreds of active AI users and remain in Stage 1 or Stage 2. Employees may complete tasks faster, but reporting remains manual, and customer information remains fragmented.

Another organisation may use fewer AI tools but have greater maturity because those tools support carefully chosen workflows with measurable outcomes.

Leaders, therefore, need to look beyond usage figures and ask:

• Which processes have changed?
• What measurable result has improved?
• How reliable is the data behind the output?
• Can the process continue without depending on one person?

These questions provide a clearer view of AI maturity than licence numbers alone.

What blocks progress?

Businesses usually stall because of operational issues rather than a shortage of AI features.

• Poor CRM data: Incomplete or inconsistent data weakens AI recommendations and forecasts.
• Unclear ownership: AI pilots lose momentum when nobody is responsible for deciding what happens next.
• Weak workflow design: Automating a poor process carries its existing problems into the new system.
• Low employee adoption: Teams need practical training on when to use AI, how to review its output and what remains their responsibility.
• No meaningful measurement: Each use case should link to an operational result, such as lead response time, forecast accuracy or reduced manual work.

Higher AI maturity requires investment beyond software fees. Businesses need time for workflow redesign, data clean-up, and change management.

Applying AI to every available task is unlikely to produce worthwhile results. Choosing the right process usually will.

Where does your business sit today?

We recommend using the five stages as a planning framework to help you decide what to do next.

At Stage 1, that means gaining visibility into current AI use and putting basic governance in place.

Stage 2 focuses on controlled experimentation, supported by suitable data and clear ownership.

Stage 3 is about making successful use cases repeatable, while Stages 4 and 5 require stronger integration, measurement and ongoing governance.

For many businesses, the most important move will be from Experimentation to Operational Adoption.

That means choosing one worthwhile revenue process, understanding how it currently works, and redesigning it with clear ownership. The team can then introduce the right AI tools, support their adoption, and measure the results.

Find out where your organisation currently sits and what your next practical step should be.