Organizations throughout sectors are present process a structural shift as AI options redefine operational workflows and problem conventional execution fashions. Software program-driven organizations are exploring how AI can increase effectivity, productiveness, creativity, and worth whereas decreasing prices. On the similar time, organizations corresponding to OpenAI, Google, Microsoft, and Anthropic are quickly releasing new variations of frontier fashions with more and more superior capabilities together with agentic options and skill to tailor to particular duties. Two years in the past, early generative AI fashions might barely full easy cyber duties. Right this moment, Claude Mythos and GPT-5.5 can autonomously execute sophisticated multi-stage assaults on weak networks. By the point this textual content is printed, new capabilities of generative AI fashions could have emerged.
This charge of change requires a shift away in engineering administration practices from conventional return‑on‑funding (ROI) calculations and remoted experimentation. Efficient AI adoption now calls for alignment of enterprise imperatives, disciplined engineering, reimagined workflows and operational processes, measurable outcomes, and steady enchancment mechanisms. Predictable readiness throughout these areas is important for preserving tempo with technical development whereas managing corresponding dangers and governance obligations.
Deliberately outlined and systematically managed AI-supported apply maturity is now a essential differentiator for organizational success. Over the previous 12 months, SEI researchers, in partnership with Accenture, have studied how organizations can mature their AI practices to convey readability, construction, and consistency to AI adoption. This publish outlines our findings on defining the scope of AI adoption, sensible steps for advancing organizational maturity, and outcomes from our joint pilot evaluation with Accenture International IT.
The Problem of Scoping AI Adoption
Many organizations are pursuing AI adoption with a imprecise, “AI all over the place” mindset reasonably than a clearly outlined technique. The pervasive AI‑washing that has saturated almost each sector—software program, telecommunications, transportation, healthcare, automotive, avionics, finance, advertising, and even small native enterprises—has served as a barrier, reasonably than an enabler. When a number of AI initiatives with totally different aims overlap with out a clear enterprise path, it turns into troublesome to prioritize worth evaluation, the supporting practices, and assets wanted.
Right this moment, AI adoption could take a number of types, every representing a significant step towards integrating AI capabilities into a corporation. Elements of AI adoption could embody
- implementing vendor options which are basically AI‑pushed (e.g., AI-assisted built-in growth environments (IDEs) or testing brokers)
- upgrading conventional vendor instruments to new AI‑enhanced variations (e.g., automated assembly summaries or AI-summarized search outcomes)
- re-imagining excessive‑worth enterprise particular use circumstances with AI-augmentation
- redesigning finish‑to‑finish workflows to embed AI elements and companies as integral capabilities that drive new methods of working
- deploying AI platforms or instruments that influence a number of workflows inside a serious enterprise perform (e.g., advertising, expertise acquisition, or contact‑middle operations)
- making an enterprise‑degree resolution to undertake AI broadly, guaranteeing the workforce is supplied, enabled, and upskilled to make use of AI wherever it provides worth (e.g., making off-the-shelf frontier fashions obtainable for common use)
The boundaries amongst these classes have successfully dissolved as a result of ease of integrating generative and agentic AI companies into enterprise environments. Importantly, ease of integration doesn’t equate to adoption maturity. Software utilization throughout the workforce is just not, in itself, proof of AI‑enabled transformation. Profitable use of a software, even by everybody within the group (e.g., a generative-AI-supported chat), is solely a software roll out. Positive aspects from software use may fluctuate. For instance, from 2023 to 2025 productiveness beneficial properties and utilization patterns of specialists shifted from a bit of to vital. AI adoption in contrast, represents a enterprise selection pushed by a strategic want with a set of targets that may be successfully addressed by an AI answer.
As AI capabilities proliferate and frontier fashions proceed to advance, organizations are starting to come across two main challenges:
1) analysis of the price of platform lock‑in, the continued operational value of sustaining and integrating AI capabilities, and the true whole value of possession related to enterprise‑scale AI adoption
2) managing a quickly evolving threat and safety panorama by which menace surfaces, assurance necessities, safeguards for delicate info, and governance expectations are regularly shifting
Moreover, organizations should confront the rising actuality that AI programs are more and more used to develop, optimize, or govern different AI programs. On this layered and extremely dynamic panorama, it’s essential for organizations to map their capabilities on to their AI adoption aims and outline scope for these efforts with precision.
The Carnegie Mellon College SEI AI Adoption Maturity Mannequin, developed in collaboration with Accenture, is designed with specific consciousness of those evolving tiers of AI use. The mannequin reinforces disciplined scope administration to deal with these challenges. Organizations that outline goal maturity ranges and institutionalize the corresponding capabilities and practices are higher positioned to use them persistently throughout numerous AI initiatives and effectively adapt them as new AI applied sciences, capabilities, and use circumstances emerge.
Growing an AI adoption maturity mannequin amid one of many quickest technological transformations in historical past presents two challenges: offering construction in a quickly evolving panorama and balancing steerage with the truth that maturity is just not a compliance train. We developed the AI Adoption Maturity Mannequin round enduring organizational capabilities reasonably than transient applied sciences, viewing maturity as a strategic selection reasonably than a prescribed vacation spot. The event of the mannequin was grounded in a disciplined, evidence-driven course of primarily based on in depth analysis, together with govt interviews, a scientific evaluate of greater than 100 present AI maturity efforts worldwide, pilots of AI initiatives, an in depth trade survey, the CMU SEI’s deep experience in maturity modeling, and Accenture’s international expertise with AI implementation. Along with the maturity mannequin, our evaluate of present AI adoption maturity fashions and frameworks will likely be launched in mid-June.
5 Steps to Scoping AI Adoption
Understanding the dedication required for AI adoption and AI maturity in at the moment’s shortly evolving technological panorama is each important and pressing. AI adoption refers back to the systematic integration of AI throughout enterprise technique, engineering practices, operational processes, and governance mechanisms. AI maturity displays the power to execute these integrations with consistency, scalability, measurable outcomes, and accountable oversight whereas adapting to fast technological and risk-related adjustments.
Till now, the instruments and applied sciences organizations used to construct software program and operationalize options for business-facing companies—although numerous—had been comparatively constant of their nature and threat profiles. Organizations as an alternative want to think about how AI-enabled initiatives will have an effect on their programs and processes throughout two essential elements:
- AI in manufacturing: the diploma of integration and autonomy of AI throughout creation and development. This side is the diploma of AI integration in manufacturing (e.g., use of IDEs) and the way independently AI-enabled programs function whereas producing options—starting from conventional to AI-assisted to augmented, semi‑autonomous, or doubtlessly totally autonomous creation (with a human within the lead in all circumstances).
- AI in system operation: pervasiveness of AI capabilities within the ensuing product or workflow. AI that will likely be in use within the ensuing workflow or product progresses equally from conventional human‑pushed merchandise to AI‑enabled, AI‑orchestrated, and autonomous programs (with a human within the lead in all circumstances). AI in programs operation could instantly help essential companies or mission-critical capabilities within the group.
Collectively, these two elements reveal a quickly rising shift within the technological panorama: organizations are getting into an area the place AI brokers are more and more relied upon for designing, deploying, and managing AI-enabled workflows and merchandise. Consequently, the excellence between the instruments organizations use (AI as growth accomplice) and the programs they construct (AI inside the workflow and product) is dissolving, underscoring the important want for steady threat administration and architectural rigor.
All programs and AI initiatives will quickly dwell within the higher left quadrant in Determine 1. One instance can be an AI-powered cybersecurity analyst the place AI brokers are used to generate artificial knowledge, develop the platform, monitor and consider the outputs, which underscores the criticality of disciplines apply, threat administration, and verification and validation. Consequently, organizations should handle dangers, governance, high quality assurance, and dependencies throughout each dimensions concurrently. This convergence is likely one of the key drivers behind the necessity for AI adoption maturity fashions as one of many devices to allow reliable programs to be developed with worth and ROI.
Determine 1: AI Adoption throughout the manufacturing and system operation axis
In the end, organizations that reach AI adoption stability pace of innovation with engineering rigor, governance self-discipline, workforce enablement, and steady studying. The AI Adoption Maturity Mannequin teams these practices into two areas of focus: organizational change and AI lifecycle engineering, with their associated dimensions and capabilities. Organizations enhance the success of their AI adoption efforts by treating AI as an organizational transformation functionality—not merely a know-how deployment. As a result of AI is now embedded each in how software program is created and the way operational programs ship worth, it have to be handled as a essential dependency and potential single level of failure. This actuality requires governance approaches that focus extra explicitly on the underlying classes of threat, not at all times controllable by customers, which the maturity mannequin helps organizations determine, assess, and handle.
To enhance the success of AI adoption efforts and obtain measurable worth outcomes amid the quickly evolving panorama of AI capabilities, leaders in control of AI adoption ought to champion the next deliberate steps:
1. Outline what AI adoption means for the group. Organizational and technical leaders typically fail to appreciate that it isn’t an AI-focused train however a business-focused train. Leaders should determine alternatives for AI to positively affect the group by answering the next questions: Why is AI wanted to realize enterprise outcomes? What areas ought to AI remodel? Organizations that fail with AI adoption don’t acknowledge that AI is a method to an finish, not the aim. The Organizational Technique Dimension of the AI Adoption Maturity Mannequin consists of functionality areas to assist organizations make progress on this regard.
2. Set a goal maturity degree that’s greatest match for the group and its targets. As illustrated in Determine 2 beneath, the AI Adoption Maturity Mannequin defines maturity throughout 5 ranges: Exploratory AI, Carried out AI, Aligned AI, Scaled AI, and Future Prepared AI. A company could select to realize a decrease degree (e.g., Carried out AI) for his or her goal maturity to align with enterprise priorities. Most organizations are prone to thrive throughout Aligned AI, Scaled AI, and Future Prepared AI ranges of maturity.
Determine 2: AI Adoption Maturity Mannequin Ranges
3. Assess your present state. Many organizations want nimble devices to information them quickly in the appropriate path and set up a staged roadmap. This isn’t a compliance train. An evidence-based, but nimble evaluation is essential. Efficient maturity enchancment requires baselining, figuring out milestones, and evaluating progress utilizing qualitative and quantitative proof. A multi-input consolidation course of consists of ongoing stakeholder engagement, metrics evaluation, tooling knowledge, artifact evaluations, and operational outcomes. A singular deal with questionnaires or static governance and compliance checks is not going to be ample.
4. Set up foundations. Organizations ought to set up core capabilities early together with governance buildings, architectural requirements, knowledge and AI lifecycle administration, measurement and monitoring practices, safety controls, and workforce coaching. Advancing AI adoption with out these foundations typically results in fragmented adoption, operational threat, and unsustainable implementations.
5. Iterate and adapt. AI applied sciences, dangers, and market situations evolve quickly. Organizations ought to undertake incremental implementation roadmaps that permit for experimentation, suggestions, recalibration, and steady enchancment whereas sustaining governance and engineering self-discipline. The ensuing evaluation approaches and roadmaps ought to allow iteration, adaptability, and evolution.
Accenture International IT Case
Placing any method to the take a look at is important in claiming dependable outcomes. We evaluated the effectiveness and use of the AI Adoption Maturity Mannequin first with Accenture’s International IT group as pilot zero.
Accenture’s International IT group serves a 786,000 international workforce and a various set of stakeholders. On the outset of the pilot, Accenture International IT demonstrated a number of foundational strengths together with a strong know-how infrastructure, a mature use case administration course of enabling fast experimentation, a coaching program, and a measurement tradition monitoring workforce-level AI utilization. In testing the AI Adoption Maturity Mannequin, our preliminary aim was to validate the mannequin in apply whereas enabling Accenture International IT to determine the subsequent frontier of AI-driven worth creation.
The pilot didn’t represent a full, formal evaluation. As a substitute, it served as an experimental validation of whether or not the AI Adoption Maturity Mannequin might precisely measure adoption and determine areas for additional enchancment even in a technologically superior group.
The pilot validated a sample in enterprise AI maturity that we had noticed in our preparatory analysis together with a survey of greater than 600 organizations performed by SEI and Accenture: organizations can exhibit a robust technical functionality whereas preserving alternatives to strengthen the structural components required to scale worth.
Challenges shared amongst these surveyed included technical deployments outpacing organizational transformation. Whereas AI programs have gotten operational, cross-functional possession, oversight, and accountability buildings are nonetheless being established. Consequently, benchmarking and price transparency are wanted to enhance ROI monitoring and funding choices. These challenges point out that whereas most organizations are transitioning from experimentation to operationalization, they’ve but to totally institutionalize the AI practices required for constant, predictable outcomes and innovation.
The pilot zero evaluation demonstrated that Accenture International IT is a high-performing group with substantial AI expertise and a robust monitor document of outcomes. Along with a evaluate of artifacts and metrics, the evaluation included interviews and workshops with totally different stakeholders reviewing the practices in opposition to the mannequin performed to cross test outcomes. On the similar time, it surfaced alternatives to extra successfully handle the complexity inherent in AI-enabled transformation each inside the group and throughout its broader ecosystem of practices to totally notice its transformative potential.
Workflow Re-engineering: AI was actively utilized to enhance workflows, together with the usage of agentic AI. Nonetheless, the evaluation recognized much more workflows that could possibly be redesigned from first ideas. In some circumstances, processes had been remodeled however lacked proof of enchancment, measurement, and standardization required to progress even additional over time.
Worth Measurement: Accenture International IT maintained a measurement tradition monitoring AI utilization, however alternatives had been recognized the place measurements could possibly be improved to seize the complete enterprise influence. By documenting value buildings inside workflows, the group might assemble rigorous ROI analyses that might evolve over time.
Governance: As an IT perform that helps the broader enterprise, the group operates inside an internet of cross-functional dependencies. The evaluation recognized a chance to additional make clear knowledge possession within the context of generative and agentic AI, outline accountability for AI failure dangers, and map dependencies — each upstream and downstream — with better precision.
These findings recognized a particular area the place organizational infrastructure might speed up the conclusion of worth from the technological adoption. Accenture International IT has a transparent aim: be a prime performer and obtain the Future Prepared AI degree of maturity. The evaluation helped them to determine concrete steps in the direction of that aim. The pilot outcomes demonstrated that the first constraint on AI maturity is alignment throughout prime precedence capabilities, corresponding to enterprise workflow innovation, measurement and evaluation, and threat and governance buildings.
The evaluation functioned as a diagnostic instrument, revealing hyperlinks that weren’t instantly seen via typical metrics. This hole represents the boundary between deploying AI and institutionalizing and bettering its influence over time.
The outcomes display that, regardless of robust technical functionality, energetic AI deployment, and robust adoption in an organizational unit, the evaluation might efficiently determine alternatives in workflow re-engineering, worth measurement, and knowledge governance that might speed up scaling AI within the group. These findings recommend that structured maturity assessments proceed to offer a dependable mechanism for diagnosing constraints in AI adoption and guiding transformation efforts. The outcomes additionally recommend that concrete practices in establishing profitable AI initiative are nonetheless evolving and these devices help in clarifying their priorities.
Classes Discovered in AI Adoption
As a part of the hassle to develop the AI Adoption Maturity Mannequin, along with Accenture International IT, we now have accomplished a number of pilots and early adopter engagements to make sure the practices in our AI Maturity Mannequin tackle essentially the most important areas in AI adoption whereas sustaining agility and readability. By way of our work on creating the mannequin and its subsequent pilots we discovered the next classes:
- Given the ever-increasing variety of AI capabilities infiltrating every thing from creation of merchandise to workflows, AI adoption maturity must be handled as a steady aim.
- AI adoption maturity assessments stay important on this more and more automation-driven panorama. Successes and failures to satisfy milestones are sometimes revealed not via written artifacts, however reasonably the unstated challenges, implicit assumptions, and omitted necessities uncovered throughout analysis and evaluation.
- As capabilities of AI companies and fashions enhance, the demand to reinvent enterprise and workflows will increase and the scope of threat shifts, placing rising emphasis on capabilities and practices that tackle threat.
As AI applied sciences, dangers, and enterprise expectations quickly evolve, organizational leaders should pursue aim alignment, steady evaluation, intentional evolution of practices, and the power to adapt governance, engineering, and operational approaches. Future posts will element patterns of gaps and roadmap priorities as we proceed to watch early-adopter engagements.
Turn out to be an early adopter of the AI Adoption Maturity Mannequin and affect the apply and evolution of AI adoption whereas getting forward of AI challenges. To be taught extra, please ship an e mail to [email protected].
To be taught extra in regards to the AI Adoption Maturity Mannequin growth journey, register for the June 9 SEI webcast the place specialists from the SEI and Accenture share technical insights and classes discovered