Organisational AI Maturity Model
In Lesson 5, you learned the four models for capturing value from domain agents. But monetisation only works if the organisation is ready for deployment. Not every organisation is.
Readiness is not a binary property -- either ready or not ready. It is a level. Organisations can move through levels deliberately, and understanding where an organisation sits today determines what intervention is appropriate. Deploy too early, and the agent fails not because the technology was wrong but because the organisation could not support it. Wait too long, and competitors move first.
This lesson gives you a five-level model for assessing any organisation's AI maturity. By the end, you will be able to look at your own organisation, diagnose its level honestly, and determine what needs to change to move forward.
Level 1: Awareness
AI is on the agenda but not in operations.
At Level 1, individual employees are using consumer AI tools -- ChatGPT, Claude, Gemini -- on their own initiative. There is no organisational sanction, no governance, no data strategy, and no designated AI owner. The leadership team talks about AI in meetings. Nobody has deployed anything.
Diagnostic Indicators
- Employees use personal AI accounts for work tasks
- No AI policy or acceptable use guidelines exist
- No budget allocated specifically for AI tools
- AI appears in strategic plans as a future initiative, not a current programme
Appropriate Intervention
Education, not deployment. Level 1 organisations are not candidates for domain agent deployment. The infrastructure -- governance, data access, designated ownership -- does not exist yet. Attempting to deploy an agent here fails not because the technology is inadequate but because the organisation cannot support it.
The right move: run awareness workshops, establish an AI working group, draft an acceptable use policy, and identify one team willing to run a pilot. That is how you move to Level 2.
Level 2: Experimentation
Active pilots. At least one team has deployed a real agent.
At Level 2, the organisation has moved beyond talk. A designated AI lead or working group exists, though with limited authority. At least one team has deployed a real agent -- not a demo, not a proof of concept, but an agent that handles real work. Results are promising but isolated.
This is where most large enterprises sit in early 2026.
Diagnostic Indicators
- At least one team has an agent in active use
- A designated AI lead or working group exists (but lacks enterprise-wide authority)
- Budget is allocated for AI experimentation (but not scaled deployment)
- Leadership is interested but not yet sponsoring enterprise-wide adoption
The Post-Pilot Trap
Level 2 is also where most enterprise AI deployments stall. The pilot worked. Leadership was impressed. And then nothing happened.
This is the Post-Pilot Trap -- the transition zone between Experimentation and Integration. Pilots succeed because they operate in controlled conditions: a motivated team, a clear problem, executive attention. Scaling requires governance, cross-team coordination, and sustained investment. Most organisations do not make that leap.
Appropriate Intervention
Deploy team-level Cowork agents with measurable value and minimal governance overhead. The goal is not enterprise transformation. The goal is building a track record of measurable results that justifies the investment needed to reach Level 3.
Pick one domain. Pick one team. Deploy one agent. Measure the value. Use that evidence to make the case for structured deployment.
Level 3: Integration
Structured deployment. Agents in production, connected to real systems, with governance.
At Level 3, the organisation has moved beyond experimentation. There is a formal AI strategy with executive sponsorship. IT has a defined role in agent deployment. Agents are connected to production systems -- CRM, ERP, document management -- with real data flowing through governed pipelines.
This is where Part 3 agents are most naturally at home.
Diagnostic Indicators
- Formal AI strategy document exists with executive sponsorship
- IT has a defined role in agent infrastructure (connectors, security, monitoring)
- At least one agent is in production with real system integrations
- Governance policies cover data access, output review, and escalation procedures
- Budget is allocated for sustained deployment, not just experimentation
Appropriate Intervention
Deploy a single vertical fully: one domain, one team, one agent, full stack. This means SKILL.md authored by the domain expert, connectors to real systems managed by IT, governance policies in place, and measurable value being tracked.
The emphasis at Level 3 is depth over breadth. Do one deployment completely and well. Document what worked, what failed, and what you would change. That documentation becomes the playbook for expanding to additional domains.
Level 4: Optimisation
Multi-vertical portfolio. Mature governance. Performance measurement driving investment decisions.
At Level 4, the organisation has multiple agents deployed across multiple domains. Governance is mature -- there are clear policies for data access, output quality, escalation, and agent retirement. The strategic question shifts from "should we deploy AI?" to "how do we optimise our AI portfolio?"
Diagnostic Indicators
- Multiple agents deployed across different departments
- Centralised governance with clear policies and oversight
- Performance dashboards tracking agent value across the portfolio
- Investment decisions driven by measured ROI, not experimentation budgets
- Platform commitment decisions (Cowork, Frontier, or both) are being made
Appropriate Intervention
This is where platform commitment and build-versus-buy decisions become relevant. At Level 4, the organisation has enough deployment experience to make informed decisions about:
- Which platform to standardise on (or whether to use both)
- Which domains to expand into next
- Where to invest in custom development versus marketplace solutions
- How to balance agent capability against governance requirements
The cross-vertical portfolio strategy in Chapter 26 is addressed primarily to Level 4 organisations. The build-versus-buy decision for SKILL.md development -- whether to invest in internal knowledge extraction capability or engage an external services provider -- becomes relevant at this level.
Level 5: Transformation
Organisational redesign around agent capability.
At Level 5, the organisation has fundamentally redesigned how it works. Job descriptions have changed. Human-agent boundaries are explicitly negotiated and documented. AI governance is a standing organisational capability, not a project.
Few organisations are here in 2026. Level 5 is the long-term destination, not a near-term goal for most Part 3 readers.
Diagnostic Indicators
- Job descriptions explicitly reference human-agent collaboration
- Organisational structure has changed to reflect agent capabilities
- AI governance is a permanent function (not a temporary project)
- New roles have been created specifically to manage human-agent workflows
- The organisation measures itself differently because of agent capabilities
Summary Table
| Level | Name | Defining Feature | Appropriate Intervention |
|---|---|---|---|
| 1 | Awareness | AI on agenda, not in operations | Education and policy |
| 2 | Experimentation | Active pilots, isolated results | Team-level Cowork deployment |
| 3 | Integration | Production agents with governance | Single vertical, full stack |
| 4 | Optimisation | Multi-vertical portfolio | Platform commitment, portfolio management |
| 5 | Transformation | Organisational redesign | Continuous evolution |
Try With AI
Use these prompts in Anthropic Cowork or your preferred AI assistant to explore these concepts further.
Prompt 1: Personal Application
I work at [describe your organisation -- size, industry, current AI
usage]. Based on the five-level Organisational AI Maturity Model, help
me assess our current level. Ask me diagnostic questions about:
(1) whether we have any AI agents in active use, (2) whether we have
a designated AI lead or governance policy, (3) whether leadership has
allocated budget for AI beyond experimentation, and (4) whether any
agents are connected to production systems. Then tell me what level
we are at and what the next step should be.
What you're learning: You are practising honest organisational assessment. The AI's questions force you to evaluate your organisation against specific diagnostic indicators rather than relying on optimistic self-assessment.
Prompt 2: Framework Analysis
Here are three organisations at different maturity levels. For each
one, identify the level and recommend the appropriate intervention.
Organisation A: A 500-person consulting firm where several consultants
use ChatGPT for research. There is no AI policy, no designated owner,
and no budget. The CEO mentioned AI at the last all-hands meeting as
"something we should explore."
Organisation B: A 2,000-person insurance company with a data science
team that built a claims-processing agent six months ago. It handles
30% of routine claims. The CTO sponsors the programme, but no other
department has deployed an agent.
Organisation C: A 10,000-person bank with AI agents in compliance,
fraud detection, customer service, and loan underwriting. A Chief AI
Officer reports to the CEO. Investment decisions are driven by
quarterly performance dashboards.
What you're learning: You are calibrating your assessment skills across the full maturity spectrum. Getting all three right confirms you can distinguish between adjacent levels (the hardest part of assessment).
Prompt 3: Domain Research
Research the current state of AI maturity in [YOUR INDUSTRY -- e.g.,
"mid-size architecture firms," "regional healthcare systems,"
"financial advisory firms"]. Based on what you find, what maturity
level would you estimate most organisations in my industry are at?
What are the most common barriers preventing them from reaching the
next level? What does that mean for me if I want to be ahead of
the curve?
What you're learning: You are positioning your own organisation within your industry's maturity landscape. This helps you set realistic expectations and identify competitive advantages available at your current level.