Energy: Where AI is scaling but controls aren't

Where AI is scaling, but controls aren't

Organizations in the energy sector are implementing artificial intelligence at a faster pace than the ability to generate the operational evidence that will be required by boards, auditors, and regulatory bodies. This is the so-called AI proof gap.

This report explains why this gap exists and what companies need to do to move from the early development phase to a model of centralized control and structured governance.

Energy companies need to strengthen AI controls

Organizations in the energy industry are not lagging behind in adopting artificial intelligence. 61% say they have already achieved measurable efficiency gains with AI, while 50% report improvements in the quality of insights for decision-making.

Grant Thornton's AI Impact 2026 survey, conducted with 950 business leaders, identified that energy companies are implementing AI at a faster rate than they are developing evidence of value and operational control mechanisms capable of meeting the requirements of boards of directors, auditors, and regulatory bodies. This is the so-called AI proof gap in the energy industry, and it is steeper than in virtually any other industry analyzed by the survey.

The consequences are concrete. Barriers related to governance and regulatory compliance are the main factor in the underperformance of AI projects in the energy industry, being pointed out by 57% of respondents, an index 11 percentage points above the average of all industries surveyed.

The boards of directors of energy companies are more involved in the oversight of AI than in much of the other industry. Still, fewer than one in five organizations are able to demonstrate that their AI controls are centralized and prepared to provide consistent evidence of compliance and performance.

Resilience is among the main drivers of AI investments in the energy industry. In this context, effective controls not only strengthen the resilience of the organization as a whole, but also increase the reliability, sustainability, and success of the AI projects themselves.

AI projects in the energy industry fail over governance

Governance barriers outweigh all other factors as a cause of poor performance of AI projects, especially in the energy industry. Governance has become the main bottleneck for the scalability of AI programs, and organizations that have not yet solved this issue are accumulating risks with each new implementation carried out without an adequate control structure.

Industry leaders have already begun to establish governance mechanisms, but the results show that there is still a long way to go. 74% of respondents say that evidence related to AI controls remains dispersed across different areas of the organization. In addition, 9% declare to have little or no confidence in the effectiveness of their controls.

At the same time, only 17% of organizations in the energy industry say they are prepared to undergo an audit of their AI governance controls and processes.

These numbers indicate that while AI adoption is advancing rapidly, the ability to prove that the technology is being used in a controlled, safe, and regulatory compliant manner still does not keep pace. Without a centralized governance structure, operational, regulatory, and reputational risks are likely to increase as new AI solutions are implemented at scale.

Governance is the biggest gap

Question: What are the top three factors that have contributed to the poor performance or failure of AI projects in your organization?

"What I observe most often is that policies exist, but they are not applied consistently or have been properly tested. The energy companies that are advancing most successfully in AI are not necessarily the ones that adopt tools the fastest. They are those that implemented controls early enough that governance never became an obstacle to deployments. In this industry, the ability to demonstrate control is what turns AI investments into organizational trust," comments Tyler Jones, Energy Sector Leader at Grant Thornton Advisors LLC.

Councils advance. The evidence does not follow up

When it comes to AI governance, the boards of directors of companies in the energy industry are among the most engaged of all the segments analyzed by the survey. This attention at the board level is important. The problem is that this involvement has not yet translated into the operational evidence required to sustain these supervisory responsibilities.

In many cases, companies still lack an operating model that can consistently connect policy, execution, and evidence. It is precisely at this point that the AI proof gap becomes most evident in the energy industry: boards of directors raise questions about liability, controls, and compliance, but operations are still unable to answer them with documented, tested, and verifiable evidence.

Councils want to participate

Question: How has your board of directors been involved in overseeing the development of the AI strategy?

"Energy industry boards are asking the right questions about AI governance. The challenge is that most organizations are still building the operational evidence needed to answer these questions," Jones reveals.

Without centralized, testable controls, governance policies and AI implementations move forward in parallel, without a clearly defined and verifiable chain of responsibility that connects one to the other.

Leaders are not confident about AI auditing

Question: How confident are you that your organization would be able to pass an independent audit of AI governance and controls (with documented evidence) within the next 90 days, including third-party and vendor-related AI risks?

Compliance is the main concern

While 30% of energy industry leaders point to regulatory changes as the top factor driving their organizations to adopt AI, 54% say regulatory or compliance uncertainty is the biggest barrier to scaling up the use of the technology.

The industry's regulatory environment creates constant tension around the advancement of AI initiatives. The decisive factor between these two realities is the existence of a governance structure capable of allowing the organization to move forward with confidence even under intense regulatory scrutiny.

Competition and regulation drive adoption

Question: What is the main external pressure driving AI adoption?

As AI outcomes move from efficiency gains in administrative activities to applications related to emissions reporting, power grid management, safety monitoring, and capital allocation, the level of control required also increases.

Organizations that have centralized, validable governance treat each new regulatory requirement as a sign of competitive advantage. Those that do not have this structure, on the other hand, end up increasing their exposure to risks.

In comparison, customer expectations appeared as the least external pressure for AI adoption in the energy industry, cited by 12% of respondents, versus 25% on average across all sectors.

In the energy industry, the main starting point for AI investments is related to operational resilience and responding to regulatory requirements. Organizations that see governance as an enabler of this response are the ones that are accelerating their adoption of technology most successfully.

Regulatory uncertainty is the main concern with Agentic AI

Question: What are your main concerns with implementing Agentic AI (Agentic AI).

Three actions to turn AI scale into success

Grant Thornton's AI Impact 2026 survey shows that boards of directors are demanding evidence. Regulators are moving forward. And Agentic AI is broadening the exposure of each implementation without proper governance. The following three actions differentiate between organizations that are able to respond to upcoming accountability demands and those that cannot.

1. Establish governance to successfully scale AI.
   57% of energy industry leaders point to governance or compliance barriers as the main cause of AI underperformance. Develop governance structures and controls that properly guide and even accelerate the successful implementation of technology.
   
   
2. Centralize control evidence before the next audit cycle.
   74% of organizations in the energy industry have fragmented controls. Fragmentation can be manageable in the pilot phase. On a corporate scale, however, it goes on to explain why leadership confidence declines and why regulatory exposure increases with each new implementation.
   
   
3. Define decision responsibilities before AI acts autonomously.
   54% of energy industry leaders point to regulatory uncertainty as their top concern regarding Agentic AI. Teams need to define which decisions AI can make independently, which require human approval, which situations require escalation, and ultimately who will be responsible for the results.

These are the capabilities that Grant Thornton helps organizations develop. This research was designed to identify the barriers to generating business value with AI, and the company supports organizations in overcoming these challenges by designing governance structures in front, middle office, and back office processes, preparing for AI compliance, aligning leadership teams, creating measurement frameworks, and preparing for the governance requirements associated with Agentic AI (Agentic AI).

The insights presented in this report reflect what Grant Thornton observes and develops with organisations on a daily basis. In today's energy market, the gap between AI capabilities and proven value is real. It is measurable. And it won't be reduced without leaders taking steps to establish verifiable governance, a solid and reliable data foundation, and the confidence to innovate.

_{Methodology
Between February 23 and March 18, 2026, Grant Thornton surveyed 950 business leaders, including CFOs, CIOs/CITOs, COOs, vice presidents, department leaders, and directors reporting directly to senior management.}

_{The energy industry-specific group was made up of 100 respondents. The conclusions related to this segment should be interpreted as indicative of the direction of the observed trends.}

_{Content Disclaimer
This content features information and commentary on current issues and developments from Grant Thornton Advisors LLC and Grant Thornton LLP. It is not a comprehensive analysis of the issues addressed. The content does not constitute and should not be construed as accounting, legal, tax, or professional advice provided by Grant Thornton Advisors LLC or Grant Thornton LLP.}

_{All relevant facts and circumstances, including the applicable technical literature, must be considered in order to reach appropriate conclusions in accordance with the topics covered in this content. For additional information on the topics covered, please contact a Grant Thornton professional.}

_{Grant Thornton LLP and Grant Thornton Advisors LLC (and their respective subsidiaries) operate under an alternative service structure in accordance with the American Institute of Certified Public Accountants (AICPA) Code of Professional Conduct and applicable laws, regulations, and professional standards.}

_{Grant Thornton LLP is a fully chartered independent public accountant (CPA) firm responsible for the provision of assurance services. Grant Thornton Advisors LLC and its subsidiaries provide tax and business advisory services. Grant Thornton Advisors LLC and its subsidiaries are not licensed CPA firms.}

_{Tax Professional Standards Statement
This content supports the professional services disclosure activities of Grant Thornton Advisors LLC and is not intended as tax advice directed at the specific facts and circumstances of any person. The content does not constitute and should not be construed as accounting, legal, tax, or professional advice provided by Grant Thornton Advisors LLC.}

_{If you are interested in the topics presented here, we recommend contacting a tax professional at Grant Thornton Advisors LLC. Nothing in this contentIt should be construed as limiting the right of any person to disclose the tax treatment or tax structure of any matter addressed.}

_{The information contained herein is of a general nature and is based on standards and guidelines subject to change. The content does not constitute and should not be construed as accounting, legal, tax, or professional advice provided by Grant Thornton Advisors LLC. This material may not be applicable or appropriate to the reader's specific circumstances or needs and may require consideration of tax and non-tax factors not described herein. Before making any decisions based on this information, consult with a tax professional at Grant Thornton Advisors LLC.}

_{Changes in tax laws or other factors may affect, prospectively or retroactively, the information contained herein. Grant Thornton Advisors LLC undertakes no obligation to inform the reader of any such changes. All references to "Section," "Sec.," or "§" refer to the Internal Revenue Code of 1986, as amended.}

_{Grant Thornton Advisors LLC and its subsidiary entities are not licensed CPA firms.}