Accenture research released alongside a Harvard Business Review article argues that companies need an integrated digital core rather than disconnected cloud, data and artificial-intelligence projects. The reported performance gap is substantial, but it describes an association among selected capabilities and results—not a promise that another technology purchase will automatically create growth.

Cloud, data and AI are presented as one capability

On 17 July 2024, Harvard Business Review published an article by Paul R. Daugherty, H. James Wilson, Karthik Narain and Prashant Shukla on building an adaptable digital core. The authors, all connected with Accenture, argued that successive waves of technology require an enterprise foundation capable of repeated change.

Accenture, a professional-services company based in the United States, defines that core broadly. It includes cloud-first infrastructure, data, AI, applications, platforms, composable integration, security and a control plane. The point is interaction: a strong model cannot help much if it cannot reach governed data or fit safely into operating systems.

The research reports a 60:40 performance gap

Accenture said its study covered 1,500 technology executives in 19 industries across 10 countries. Companies combining an advanced digital core, greater strategic-innovation investment and a balanced approach to technical debt recorded a 60% higher revenue-growth rate and 40% higher profitability than comparison companies.

Those figures need careful reading. The release reports a relationship in Accenture's research; it does not describe a randomised experiment proving that the three practices alone caused every difference. Management quality, market position, capital and earlier investment may influence both digital maturity and performance. The findings are a useful benchmark and hypothesis, not a guaranteed return.

The core contains several mutually dependent layers

  • Cloud-first infrastructure that can scale and change.
  • Governed data and AI capabilities tied to business decisions.
  • Applications and platforms that support actual workflows.
  • Composable integration instead of brittle point-to-point connections.
  • Security and operational control across every layer.

Innovation cannot outrun the foundation forever

A company can launch an AI pilot quickly by connecting a model to a narrow dataset. Scaling it across finance, customer service or production is harder. Identity, data definitions, latency, audit evidence, cost limits and ownership become operational questions. If each project creates another isolated stack, apparent speed today becomes integration work tomorrow.

A coherent core does not mean one giant system or one vendor. It means agreed interfaces, shared controls and reusable services. Teams should be able to add a new capability without rebuilding authentication, data movement, monitoring and recovery every time. Modularity makes change safer only when the modules connect under clear rules.

A text-free physical infographic joins layered cloud, data, application and security foundations with gated innovation experiments and the sorting of tangled legacy technology into two rising performance tracks
The research combines three disciplines: strengthen the core, fund strategic innovation and keep accumulated technology obligations manageable.

AI is also becoming a source of technical debt

Accenture reported that 41% of surveyed executives ranked AI among their three largest contributors to technical debt, tied with applications and platforms. That result complicates the idea that AI merely modernises old systems. Hastily selected models, duplicated data pipelines, unmanaged prompts and opaque dependencies can create a new maintenance burden.

Technical debt is not automatically bad. A deliberate shortcut can help test demand before committing to a permanent architecture. The danger arises when temporary components lose an owner, documentation or retirement date. A balanced portfolio identifies which debt buys learning, which supports a critical legacy process and which only consumes capacity.

Leaders need evidence at the level of business flow

An executive team should not judge the core by cloud migration percentages or the number of AI pilots alone. Better measures follow a business change from idea to result: time to expose trustworthy data, time to integrate a new service, frequency of safe releases, recovery time, reuse of common capabilities and the cost of maintaining old components.

Accenture's July research gives companies a useful framing for technology investment. Its strongest implication is that adaptability cannot be bought as an isolated product. It emerges when architecture, data, security, operating responsibility and disciplined retirement of old technology allow useful innovations to move repeatedly from experiment into reliable work.