AI Adoption Maturity Model v1.0 (CMU/SEI)
In 2026 the Software Engineering Institute at Carnegie Mellon University released the “AI Adoption Maturity Model v1.0 – Transforming Organizations Through AI.” [https://www.sei.cmu.edu/documents/6534/The_AI_Adoption_Maturity_Model.pdf] This 5-Level framework (with Capability Areas, Goals, Practices, etc.) is structurally similar to the CMMI and is designed to help organizations progressively migrate from their current state to greater levels of successful AI understanding, adoption, deployment, and application.
The following Table is a cut-paste copy directly from section 4.14 of the “AI Adoption Maturity Model v1.0” (Carnegie Mellon University, Software Engineering Institute)…
If your first impression is that 25 AI Adoption Capability Areas is a lot to think about all at once, you are absolutely correct. Abridge Technology has designed a suggested implementation sequence that prioritizes early success, reliable progress, and a progressively stronger foundation of AI capability. This 7-phase adoption sequence [using material from the AI Adoption Maturity Model v1.0, CMU/SEI] is outlined below.
AI Adoption Phases
This phased approach provides a sequence of focus efforts so that you are not ‘focusing’ on everything at once. The left column on the Table below contains the Section number and Title from the “AI Adoption Maturity Model”. The columns to the right, labeled P1 through P7, are the implementation Phases (described below) recommended for sequencing your efforts and progressively continuing your transition into the AI world.
Focus Capability Areas
In the material below, the number leading the Capability Area name is the Phase sequencing number. The number in [braces] trailing the Capability Area name is the Section number where you can find details from the “AI Adoption Maturity Model v1.0 – Transforming Organizations Through AI.” [https://www.sei.cmu.edu/documents/6534/The_AI_Adoption_Maturity_Model.pdf]
Phase 1 focus Capability Areas:
- 1.1 AI-Literate Workforce Development Capability Area [4.5.1]
- 1.2 Experimentation Capability Area [4.6.1]
- 1.3 Risk Management Capability Area [4.7.1]
Rationale for Phase 1 focus: The objective of this initial step is to ensure that designated personnel are adequately AI-literate so they can successfully experiment with potentially useful AI tools and techniques. Since this phase involves a fair amount of risk, the Risk Management Area is likewise an early priority for nearly all organizations.
Phase 2 focus Capability Areas:
- 2.1 AI Architecting Capability Area [4.9.1]
- 2.2 Test and Evaluation Capability Area [4.9.2]
- 2.3 Data Lifecycle Management Capability Area [4.8.1]
- 2.4 Data Quality Assurance Capability Area [4.8.2]
Rational for Phase 2 focus: Once the workforce is AI-literate, has commenced with experimenting, and is conducting active risk management, the focus can expand to includes AI architecting, testing, and evaluation. With AI in particular, the quality of the results you achieve will highly correspond with the quality of the data you are using for training and/or inputs. Hence, both data lifecycle management and data quality assurance become essential to success.
Phase 3 focus Capability Areas:
- 3.1 Strategy Development Capability Area [4.4.1]
- 3.2 AI Model and Agent Security Capability Area [4.10.2]
- 3.3 Responsible AI Capability Area [4.7.3]
- 3.4 Policy and Compliance Capability Area [4.7.2]
Rationale for Phase 3 focus: It could be easily argued that AI Model and Agent security should be in an earlier focus area than presented here (and in highly secure environments this is almost certainly true). However, from a different perspective it could be rather difficult to establish effective security prior to knowing what it is you are securing. Both Phase 1 and Phase 2 help lay a strong foundation for you to effectively evaluate and implement reliable security practices. Parallel to this, you have enough in place for establishing an overall strategy, supporting policies, and compliance practices. Central to all this is ensuring responsible AI capability.
Phase 4 focus Capability Areas:
- 4.1 Technology Infrastructure Capability Area [4.11.1]
- 4.2 AI Model Management Capability Area [4.10.1]
- 4.3 Human-in-the-Lead Automation Capability Area [4.6.3]
- 4.4 Business Workflow Innovation Capability Area [4.6.2]
- 4.5 Measurement and Analysis Capability Area [4.6.4]
Rationale for Phase 4 focus: This is potentially the ‘heaviest lift’ of any of the Phases due in part to being the only Phase where you are focusing on 5 new Capability Areas. However, the establishment of these Capabilities likely has a greater impact on the management levels of the organization than on the engineering and support levels. Also, infrastructure, AI model management, business work-flow, and human-in-the-lead activities all potentially highly impact each other, and therefore might best be approached by taking a deliberately integrated perspective. As these areas are clarified, and practices established, it likewise becomes imperative that you establish complementary and supportive measurement and analysis practices to better provide transparency and timely insights into whether or not your measurement goals (and corresponding business goals) are being achieved–and at what rate.
Phase 5 focus Capability Areas:
- 5.1 Monitoring Capability Area [4.10.3]
- 5.2 Organizational Culture Evolution Capability Area [4.5.2]
- 5.3 Legacy Integration Capability Area [4.9.3]
- 5.4 Transparency and Explainability Capability Area [4.9.4]
Rationale for Phase 5 focus: At this Phase the focus expands to better ensure that you are integrating AI practices and tools into your legacy environment and systems, and that you are correspondingly supporting the evolution of your business culture to ensure compatibility between behaviors, tools, and practices. This can be achieved through monitoring key indicators, ensuring transparency, and being able to effectively and accurately explain what you are seeing, and why you are seeing it.
Phase 6 focus Capability Areas:
- 6.1 Partnership Development Capability Area [4.4.2]
- 6.2 Supply Chain Management Capability Area [4.11.2]
- 6.3 Organizational Structure Calibration Capability Area [4.4.3]
Rationale for Phase 6 focus: As AI tools, methods, and environments become increasingly stable components of your environment, you can more reliably expand your focus beyond the internal environment and start enhancing your capabilities relative to key relationships and activities outside your environment. This includes establishing and managing partnerships and ensuring adequate, appropriate, and secure involvement of the providers that are part of your supply chain.
Phase 7 focus Capability Areas:
- 7.1 Deployment at Scale Capability Area [4.11.3]
- 7.2 Future Ready Planning Capability Area [4.4.4]
Rationale for Phase 7 focus: These two Capability Areas are the only two Areas that are portrayed at Maturity Level 4 and Maturity Level 5 of the AI Adoption Maturity Model (CMU/SEI). As such, it is likely you want as many other Capability Areas as possible to be effectively established and efficiently operating. With such a foundation you can deploy at scale with correspondingly less risk. Additionally, and especially given the stunning rate at which the AI world is advancing, you can, with increased clarity and confidence, establish and deploy future-ready planning for whatever is coming next.
Summary: AI Adoption Focus Phases
Of course, the above will be highly influenced by many different factors that are specific to your organization. Key questions to consider include:
- To what degree have you already adopted AI techniques, tools, and methods?
- What type of “AI organization” are you? At one end of the spectrum, you use AI for internal efficiency. Alternatively, you may provide AI as part of your solution set. Or maybe you are building and training AI engines? Maybe your organization is a mix of all of these?
- What level of transparency–supported by objective measurements–do you currently have relative to your adoption of AI?
- What has been your AI adoption success rate and scope?
- How sensitive is the data you are handling?
- How thoroughly do your current security techniques address the security threats associated with AI adoption?
Your answers to the above (and numerous other) questions may well motivate you to adjust the sequencing of your approach to covering the Goals and Practices in the 25 Capability Areas of the AI Adoption Maturity Model (CMU/SEI). In any event, it is highly recommended that you take steps to avoid trying to do everything at once. By definition, when everything is a top priority, you have no top priority.
Feel free to contact Abridge Technology (rbechtold@abridge-tech.com) if you have any questions about any of the above.