Insights from Chris Dimitriadis, Chief Global Strategy Officer, ISACA.getty​The capabilities of AI to accelerate innovation continue to captivate enterprise leaders, but those aspirations and the reality often are mismatched. The massive investment AI is creating will only be sustainable if AI projects deliver real results, and it's still, in many cases, struggling to clear that bar. ISACA's 2026 AI Pulse Poll surveyed more than 3,400 digital trust professionals and found that return on investment (ROI) around AI is still a work in progress. Only 22% of the global respondents indicated that AI ROI has met or exceeded expectations, vastly outweighed by the combined responses of those who indicated the ROI has been lacking, don't know or say it's too early to tell.This leaves boards of directors and enterprise leaders in an uncomfortable position, especially when combined with AI's volatile relationship with business imperatives such as performance, risk, trust and reliability.This isn't a technology issue. AI's transformative capabilities are already evident, with compelling use cases on display across industry sectors. Rather, much of this untapped potential can be attributed to AI governance gaps and an insufficiently structured path for organizations to effectively and responsibly integrate it. Achieving sustainable success with AI doesn't have shortcuts, but with a sturdy governance foundation, delivering on the ambitious goals to which many organizations aspire has a realistic path.Those goals exist for good reason, as the potential for transformative innovation through AI is real in the form of modernized business models, enhanced efficiency and more scalable product development. Structured governance should be viewed as an enabler to this desired innovation, not a roadblock.Effective governance ensures that AI initiatives align with overall business strategy and objectives—a common disconnect for organizations that haven't realized the benefits from AI they anticipated. Governance also standardizes processes, which makes it easier to efficiently execute complex AI implementations. Additionally, this structured approach should include putting a performance measurement in place that keeps track of the effectiveness of AI initiatives and provides timely input to change course as needed.By ensuring AI initiatives are strategically aligned, well-resourced and effectively monitored on an ongoing basis, structured governance enables organizations to harness the full potential of AI technologies with a constructive balance of innovation alongside effective oversight.Taking a structured approach is especially important given the complexity of AI systems and implementations. As AI models grow increasingly sophisticated, they can be opaque and difficult to comprehend, even for their creators. This can lead to unpredictable behavior, biased algorithms and ethical shortcomings—all of which can lead to harmful outputs with the potential to badly undermine customer trust.Another layer of complexity comes into play as enterprises engage with vendors using AI, which oftentimes brings additional risk about that (absent strong AI governance) organizations may be unaware of. On a fundamental level, AI's probabilistic reasoning introduces uncertainties. A United Nations University article noted that "recognizing this reality is crucial for developing governance frameworks prioritizing resilience, adaptability and ethical safeguards."Other critical governance needs that are closely coupled with ROI from AI investments relate to the need to redesign processes rather than just introduce AI into existing ones, making sure that data structures are ready for AI processing and upskilling employees for ensuring readiness of the workforce to design, build or acquire, monitor and improve AI implementations. The triangle of process, people and data/technology is key for ensuring AI projects have a platform to produce a ROI in a measurable and controlled manner.Given these challenges, a structured governance approach to AI is a necessity that must be emphasized from the board level, with sharp organizational focus on control, traceability and accountability in AI implementations. This structured approach can be enabled through leveraging industry frameworks, especially frameworks that have been designed or updated to include AI-specific guidance. We created our new framework, the CMMI Artificial Intelligence Maturity (AIM), through the domain lenses of data, development, people, safety, security, services, suppliers and virtual to help ensure AI initiatives extend beyond technical implementation capabilities to critical enablers such as workforce readiness, data governance and ethical considerations.Beyond frameworks, a traditional approach to ROI would relate to identifying metrics of impact from AI implementations, trying to quantify productivity increase (man hours reduced in workflows), increase in customer satisfaction, increase of sales in products that introduce AI or improved resolution times in customer calls. Overall, ROI calculation should be approached in the context of the business (what matters most).​To realize the intended benefits of AI, organizations need to manage related risks, account for regulatory and ethical obligations, strengthen guardrails and work toward a pathway for sustainable AI-enabled performance improvement. This can only reliably be achieved through a disciplined approach to governance that enables organizations not only to think big about AI, but to operationalize it in a way that justifies their efforts and investments. By prioritizing robust governance and leveraging relevant frameworks, enterprises can spark responsible innovation while bridging the gap between AI aspirations and reality.Forbes Technology Council is an invitation-only community for world-class CIOs, CTOs and technology executives. Do I qualify?