NTT Data: Beyond Infrastructure: Cloud-Led Innovation in the AI Era

Artificial intelligence is rapidly reshaping how enterprises operate and create value. As AI transitions from experimental initiatives to large-scale, AI-driven execution, the cloud’s role has fundamentally evolved. It is no longer merely an enabling technology; it has become the critical execution layer for AI operating models, enabling AI systems to reason, act, and scale effectively. Yet, despite nearly two decades of cloud adoption, many organizations have not achieved the necessary levels of cloud maturity to fully capitalize on AI’s potential.

Cloud-led innovation in the era of AI, a global report by NTT DATA, based on a survey of over 2,300 senior decision-makers across various industries, highlights a significant gap between cloud ambition and reality. The research indicates that organizations failing to evolve their cloud foundations risk constraining the growth and value of their AI investments. “Cloud leaders”—organizations identified as “cloud-evolved” due to advanced cloud adoption, strong revenue growth, and high operating profit—are nearly 2.5 times more likely to deliver revenue growth exceeding 10% and more than twice as likely to operate at margins of 15% or higher, compared to their peers. These leaders adopt a distinct approach to cloud, treating it as a strategic foundation for long-term innovation rather than solely an infrastructure platform.

Strategic Alignment and Architectural Foundations for AI

To effectively harness AI, enterprises must move beyond viewing cloud as a purely technical consideration. Success in the AI era demands a deliberate, integrated strategy that aligns cloud capabilities with business objectives and robust architectural foundations.

Develop Cloud and AI Strategies in Tandem

The report underscores that cloud and AI strategies are now inseparable. Cloud is the operational core where AI models make decisions, trigger actions, and scale efficiently. When these strategies are developed in isolation, organizations risk misalignment between their AI ambitions and the underlying technological execution capabilities. A critical challenge identified is a shortage of AI skills within the cloud context, with 49% of cloud leaders anticipating this will remain a top skills challenge over the next 12 months. This highlights the need for a unified approach to talent development and capability building.

Furthermore, the influence of Chief AI Officers (CAIOs) is growing, with 61% of CAIOs strongly agreeing that the rise of agentic AI has increased the need for cloud investment, compared to 50% of CIOs/CTOs . This indicates a clear demand for integrated strategic oversight at the executive level.

What to do:

• Establish a joint Cloud and AI steering committee involving the CAIO, CIO, CTO, and relevant business unit heads. Define shared objectives and key results (OKRs) for AI-driven cloud initiatives.
• Develop a unified talent strategy that addresses the AI skills gap within cloud engineering and operations teams, potentially through targeted training programs focusing on MLOps, cloud-native AI services, and data engineering for large language models (LLMs).
• Integrate AI readiness into cloud roadmaps. For instance, when planning the migration of a customer relationship management (CRM) system, ensure the chosen cloud architecture and services support future AI integration for personalized CX or intelligent automation of customer service workflows.

Strategic Cloud Architecture Choices Drive AI Success

Cloud deployment choices directly influence AI outcomes, impacting how AI workloads scale, perform, and are governed. Decisions regarding landing zones—preconfigured, secure, and scalable environments—are crucial. The report highlights a growing interest in private and sovereign cloud models, with 99% of organizations expecting private cloud adoption to increase due to concerns about data sovereignty, ownership, cost, and security. Sovereign cloud adoption is projected to rise 50% in the next two years, from 28% to 42% of organizations. This shift is driven by requirements for data security, privacy, and compliance, especially for sensitive data and regulated workloads.

AI workloads are often computing-intensive, data-hungry, and unpredictable. Distributed cloud environments—hybrid and multicloud setups—introduce complexity that necessitates robust governance embedded directly into the architecture. This includes standardized architectures, shared identity frameworks, policy defined as code, and automated enforcement mechanisms.

What to do:

• Conduct comprehensive architecture assessments before new AI initiatives, specifically evaluating data residency, security, and performance requirements for AI workloads.
• Prioritize private or sovereign cloud deployments for AI systems handling sensitive customer data or intellectual property (e.g., a financial services firm developing AI for anti-money laundering, requiring data to reside within national borders for regulatory compliance).
• Implement policy-as-code for cloud resource provisioning and AI model deployment, ensuring guardrails for cost (e.g., GPU usage thresholds), security (e.g., data encryption at rest and in transit for all AI data pipelines), and compliance are enforced automatically (e.g., GDPR, HIPAA).
• Define clear data governance policies for AI data, including data lineage, access controls (role-based access control, attribute-based access control), and consent management, ensuring alignment with customer trust expectations.

What to avoid:

• Deferring architectural decisions or treating cloud as a generic hosting environment without considering AI-specific needs for scale, data proximity, and regulatory adherence.
• Adopting cloud-neutral strategies that fail to differentiate between workloads with varying data sensitivity or performance requirements.

Section 2: Operationalizing AI Value Through Platforms and Metrics

Translating AI capabilities into tangible business value requires robust operating models, underpinned by platform-led approaches and outcome-driven measurement.

A Platform-Led Operating Model is Essential

The complexity introduced by distributed cloud environments and AI-driven workflows necessitates a shift from manual, team-dependent management to a technology-driven, platform-led operating model. Organizations expect their use of fully managed and automated cloud platforms delivered by strategic partners to triple over the next two years, from 11% to 32%. These platforms are crucial for orchestrating, monitoring, and enforcing guardrails across the entire technology stack, enabling control without impeding innovation.

A platform-led approach delivers end-to-end visibility of how business processes execute across systems, data, integrations, and AI-driven actions. This allows leaders to gain real-time insights into performance, risk, and value . The industrialization of platform operations is evident, with nearly half of organizations (49%) executing Site Reliability Engineering (SRE) practices entirely through a service provider.

Operating Model and Roles:

• Establish a dedicated Cloud Platform Engineering team responsible for designing, building, and maintaining standardized cloud environments, MLOps pipelines, and shared services (e.g., data lakes, API gateways).
• Define clear Service Level Agreements (SLAs) with cloud providers and managed service partners, specifically addressing uptime for AI inference engines (e.g., 99.99%), data processing latency (e.g., <100ms for real-time recommendations), and incident response times.
• Implement a robust observability framework integrating logs, metrics, and traces from cloud infrastructure, AI services (e.g., prompt engineering, function calling), and business applications into a unified platform (e.g., Datadog, Splunk), providing a single pane of glass for operational health and business impact.

What to do:

• Consolidate and standardize cloud automation and productivity efforts into a single, platform-based cloud management approach.
• Automate routine operational decisions and codify best practices within the platform to ensure consistent, scalable operations for AI workloads.

Reset Cloud Transformation KPIs for Business Value

Traditional cloud transformation KPIs, such as the number of applications migrated or infrastructure cost savings, track activity rather than true business value. In the dynamic, AI-driven cloud environment, these static, backward-looking metrics are insufficient. The report highlights that 62% of organizations expect the impact and ROI of agentic AI for cloud-native application modernization to fall short of expectations. This suggests a disconnect in how success is measured.

AI fundamentally changes what is possible during migration and modernization, enabling automation of workload discovery, dependency mapping, and performance simulation. This compression of transformation timelines requires continuous, value-based, forward-looking measurement. KPIs must shift from “did we migrate?” to “are modernized environments behaving as expected, are costs predictable, and are AI-enabled workflows improving business outcomes?”.

What ‘good’ looks like:

• Shift from technical metrics to business outcomes. For a retail e-commerce platform, instead of measuring “VMs migrated,” track “increase in AI-driven personalized offer conversion rate” (target: 5-8% increase), “reduction in cart abandonment due to AI-powered recommendations” (target: 10-15% reduction), or “improvement in customer satisfaction (CSAT) for AI-assisted support interactions” (target: 0.5-point increase).
• Implement real-time dashboards that link cloud resource consumption and AI model performance directly to business metrics like customer retention, revenue growth, and operational efficiency. For a B2B SaaS provider, this might involve tracking “time-to-resolution” for AI-powered IT service management (target: 30% reduction from 2 hours to 1.4 hours) or “first contact resolution (FCR) rate” for AI-driven customer support (target: 10-15% improvement).
• Integrate AI-driven insights into KPI reporting. Use AI to analyze performance data and predict potential cost overruns or performance bottlenecks, allowing for proactive adjustments.

Immediate Priorities (First 90 Days):

• Review existing cloud KPIs and identify those that are purely technical or activity-based.
• Convene business and IT stakeholders to redefine KPIs that directly measure the impact of cloud and AI investments on strategic business objectives.
• Pilot new value-based metrics on a small but impactful AI initiative, demonstrating the shift to outcome-oriented measurement.

Securing the AI-Driven Cloud

As cloud and AI ecosystems become more distributed and complex, security is no longer an afterthought but a business-defining priority. It is the number-one investment priority for cloud.

Make Cloud Secure with a Focus on the Basics

The integration of AI introduces new attack surfaces, from complex data pipelines feeding AI models to automated decision paths. Security, governance, risk, and compliance concerns related to autonomous agents are the top challenges for AI adoption in cloud-based solutions over the next 12 to 18 months. Legacy lift-and-shift migrations often carry existing security gaps into the cloud, where they can be amplified.

Security cannot be managed in isolation; it requires an enterprise-wide view of risk that spans providers, platforms, and partners. This necessitates built-in accountability and governance. While hyperscalers offer robust native security capabilities, organizations must assume ownership of risk and avoid relying solely on a single provider’s controls. Only 27% of cloud leaders rely primarily on a cloud provider’s native security controls, compared with 44% of other organizations.

Governance and Risk Controls:

• Implement a robust identity and access management (IAM) framework across all cloud environments, integrating with enterprise directories and enforcing least privilege access for both human and AI service accounts.
• Define clear data protection policies for AI data, including encryption at rest and in transit, data loss prevention (DLP) for sensitive information used by AI models, and data retention schedules.
• Establish a proactive risk management and governance framework for AI, including regular threat modeling of AI architectures, red-teaming for AI models to identify vulnerabilities (e.g., prompt injection, model inversion attacks), and a clear incident response plan specifically for AI-related security breaches.
• Mandate continuous monitoring for cloud and AI environments using Security Information and Event Management (SIEM) and Extended Detection and Response (XDR) platforms, with automated alerts for anomalous AI model behavior or unauthorized data access.
• Ensure clear delineation of shared responsibility for security between the enterprise and its cloud managed service providers, documenting roles and responsibilities in formal agreements (e.g., for patch management, vulnerability scanning of underlying infrastructure).

What to do:

• Adopt a top-down, enterprise-wide approach to security that embeds strong identity management, clear data protection policies, proactive risk management, and continuous monitoring into cloud platforms and operating models.
• Embrace private and sovereign cloud solutions for AI workloads that handle highly sensitive or regulated data, providing enhanced control over data location and access.

What to avoid:

• Treating security as an afterthought or a checkbox exercise; it is a prerequisite for realizing value from AI.
• Outsourcing complete security responsibility to hyperscalers without maintaining internal oversight and tailored controls.

Summary

The convergence of cloud and AI marks a pivotal moment for enterprises. The NTT DATA report “Cloud-led innovation in the era of AI” (2026) unequivocally states that cloud is no longer just where systems run; it is the execution layer where AI operates, decisions are made, and value is created at scale. Organizations that embrace this reality and adapt their cloud strategies accordingly will be better positioned to compete and innovate.

Achieving this requires a holistic transformation: aligning cloud and AI strategies in tandem, making deliberate architectural choices, reimagining business value through modern applications and data readiness, adopting a platform-led operating model, resetting KPIs to focus on business outcomes, and fundamentally securing the entire AI-driven cloud ecosystem. By proactively addressing these imperatives, senior marketing and CX leaders can ensure their cloud investments deliver sustained business value and competitive advantage in the AI era.