Despite making massive investments to implement AI, most organizations are still struggling to get meaningful value from it. The problem is rarely the technology or the data. Those are important, but a bigger issue is that companies often don’t devote enough attention to leveraging people, processes, and the overall operating model. In BCG’s experience, just 10% of the value from AI comes from algorithms and 20% from data, while the remaining 70% comes from changes to the operating model and new ways of working. To address this issue and make faster progress, companies should consolidate their AI efforts and resources in an AI hub.
Similar to a center of excellence, an AI hub helps companies turn scattered AI activity into repeatable business value. It aligns the company’s entire portfolio of initiatives around a clear mandate, supported by shared data foundations, strong delivery discipline, and governance.
AI hubs can be extremely flexible in design, and they can—and must—evolve with the changing needs of the organization. In most cases, they can help companies be more deliberate about org design so that they can capitalize on the potential from AI.
Lots of Activity, Little in Terms of Results
AI is unlike other technology developments in that it cuts across functions, workflows, data domains, risk boundaries, talent models, and decision rights. One recent BCG analysis found that 73% of CEOs report that they have the final say on AI decisions, up two-fold from last year. AI forces teams in different parts of the company—technology, business, finance, risk, HR, and legal—to move in a much tighter loop than most organizations are designed to support. And it changes how work gets done, who does it, and how leaders measure performance.
That is why many companies are struggling to capture value from AI. In BCG’s 2026 AI at Work survey, half of respondents said their companies lack clear governance for managing teams with people and AI, and almost as many said that AI-related accountability is one of their three top concerns for the future. Many leaders have probably experienced the challenges firsthand. Business units identify use cases and launch pilots, but few scale up across the enterprise. Technology teams deploy platforms and tools without HR addressing how ways of working will change. Risk, legal, and cyber teams define guardrails and governance, but they’re not properly embedded in project delivery processes, and senior leaders see lots of activity but little in the way of meaningful results.
The good news? For the companies that improve their AI performance, the returns are significant. Future-built companies have 1.7 times the revenue growth of laggards, and 3.6 times the three-year total shareholder return. That difference comes from focusing a company’s resources to turn AI investment into sustained advantage. An AI hub can be a critical way for other companies to start capturing AI value and ultimately join this group of leaders.
What the AI Hub Does
An AI hub is a dedicated entity with the mandate to coordinate all AI initiatives throughout the organization. It can comprise as few as 10 full-time employees or as many as 100 or more, reflecting the flexibility of the approach. It can house critical capabilities or link them together across the enterprise, and—as we discuss below—the structure can change over time.
No matter the size, all AI hubs must have cross-functional expertise and explicit authority to oversee all of the company’s AI efforts. They serve four essential functions:
- Coherence. The AI hub sets the overall strategic direction for AI. It aligns the portfolio of initiatives across business units and functions, while also overseeing architecture, data, tooling, standards, and investment priorities. This prevents duplication, concentrates scarce talent, and helps the organization build capabilities that can be reused in multiple use cases.
- Speed. The hub accelerates projects by bringing together cross-functional expertise across product management, engineering, data science, business expertise, and change capabilities. In that way, it develops repeatable assets, methods, and platforms that become more effective over time.
- Governance. The hub embeds governance. Responsible AI, cyber, legal, risk, compliance, and value measurement get integrated at the start of every project, spanning design, deployment, monitoring, and scaling. The hub makes governance operational, so controls are built into how AI moves through the enterprise.
- Adoption. The hub drives adoption to change ways of working among employees and managers so that new behaviors take root and teams become more AI-capable over time. That means going beyond technology deployment to focus on change management, behavioral science, and workforce redesign.
In short, the hub serves as a mechanism for ensuring that AI strategy, technology, risk, talent, and behavior change move together.
Where the AI Hub Should Sit
There is no one-size-fits-all answer to where the AI hub should sit in the organization; the decision on reporting lines must be context-specific. It can sit under the CIO or CTO to ensure tight integration with enterprise technology. It can report to the chief strategy officer to reinforce strategic alignment. It can sit under the CFO to strengthen capital allocation and the tracking of value for AI initiatives. It can report to the chief HR officer to support redesigning the relationship between the human and digital workforce. In some companies, it should report directly to the CEO to provide maximum cross-functional authority.
However, there are a few universally applicable considerations that should determine the best approach for each organization. The right answer depends on what the company believes AI represents: a technology program, a productivity agenda, a workforce transformation, a strategic growth engine, or a redesign of the enterprise operating model. But the decision itself must be made at the CEO level.
How the AI Hub Evolves
The structure and main objectives of an AI hub are not fixed. Instead, they evolve as the enterprise grows in AI maturity, with projects and discipline typically centralized in the hub early on and shifting to business units as teams become more capable. (See “Key Leadership Decisions.”)
Key Leadership Decisions
- Who’s in charge? The first job is to appoint and empower the right AI hub leader. The best hub leaders are systems thinkers. They combine technology fluency with commercial acumen, they understand how large organizations actually change, and they can disrupt the status quo and work across functions without turning AI into a turf battle.
Often, AI hub leaders are visible candidates in CEO succession planning. AI is among the highest-profile efforts at most organizations, and it entails knowing how value is created throughout the enterprise, how decisions get made, and how to build momentum across functions. - What’s the biggest stumbling block? The second job is to diagnose where AI efforts are currently falling short. Most AI programs stall because of unclear ownership, insufficient adoption, unforeseen risks, or unmeasured value. Leadership teams should determine which of these areas are the biggest priority and shape the hub’s mandate accordingly.
- What’s our starting point? The third job is to choose the right starting point. Starting in a highly decentralized manner creates fragmentation. Waiting too long to distribute ownership creates a central bottleneck. Many companies should begin with a hub-led model and evolve toward greater business unit ownership over time. Others may already be mature enough to start closer to phase 2. Either way, the path should be deliberate and tailored to a company’s objective needs.
As the exhibit shows, there are four main phases in a hub’s development.
Phase 1: Hub-Led. In the first phase of maturity, the AI hub implements and delivers AI initiatives. Business units identify priorities and provide domain expertise, but the hub is the source of virtually all AI talent. The goal at this point is to demonstrate value quickly and credibly, with the company focusing its scarce resources and building the discipline needed to deploy AI responsibly and capture impact.
For example, a medtech company created an AI hub, starting with approximately 15 full-time employees, to accelerate the delivery of initial use cases. By bringing together portfolio strategy, a foundational AI platform team, risk assessment, and project execution, the company was able to shorten the time to market for new use cases by 50%.
Phase 2: Center and Pod. In the second phase, the focus shifts from proving AI value to building the institutional muscles to scale it. The AI hub still provides significant support, but the business units begin to lead more use cases, sharing AI resources with the hub.
This is where the people agenda becomes critical. Business units need leaders who can sponsor AI-enabled change, teams that can work in new cross-functional ways, and employees who willingly adopt AI into daily workflows. The hub continues to provide standards, governance, and expert support, while helping develop and embed AI capabilities across the company.
A consumer goods company followed this path by launching an AI transformation in marketing and then expanding to other functions. An enterprise AI council set the overall direction, centralized AI delivery sat within the CIO’s organization, and delivery squads operated between IT and the business, drawing in subject matter experts from the business units. Starting with one high-visibility function helped the organization prove the value from AI, build reusable components, and create momentum before expanding more broadly.
Phase 3: Business Unit-Led. In the third phase, the AI hub becomes a strategic center of excellence, focused on monitoring, enabling, governing, and continuously improving the enterprise AI system. Business units own and deliver AI use cases.
To succeed in this phase, companies need AI capabilities and tools embedded across the organization, including a modular shared data platform and sufficient AI talent in the business units. Mature risk and cyber controls must be built into platforms and workflows. Measurement mechanisms need to be established and reliable. Most important, the company needs business leaders who genuinely support the digital agenda and can develop the AI talent they inherit. When those conditions have been met, AI ownership can move closer to the business without fragmenting the enterprise approach.
A large technology company moving toward this model has approximately 100 full-time employees in its AI organization, with most resources sitting in the business units while the hub governs and supports the enterprise platform model.
Phase 4: The AI-First Organization. The AI hub is the engine that helps companies scale AI today. Over time, it may also become the bridge to a more AI-first organization.
In an AI-first operating model, AI agents increasingly orchestrate work that previously sat within the business units. The boundary between the hub and the business begins to blur because AI systems operate across functions, processes, and decision flows. The hub’s role evolves from enabling AI delivery to helping operate and govern a digital and human workforce working together.
Of course, the AI-first organization is a long-term objective for most organizations. But it matters now because the choices leaders make today will determine how ready their companies are for more agentic ways of working. Those that build strong AI hubs will be better positioned to decide where AI should augment human work, where it should automate work, and where it should reshape the enterprise itself.
AI is becoming ubiquitous, and every company has access to powerful models, cloud infrastructure, vendors, and tools. These inputs matter, but they are not the ultimate differentiator. The real advantage comes from the operating model: the structure, governance, leadership, talent, and behaviors that allow AI value to compound.
For chief HR officers, the implications are especially significant. As AI moves from experimentation to scaled adoption, the hardest questions increasingly concern roles, skills, leadership behaviors, organization design, workforce trust, and how human and agentic work will coexist. The chief HR officer is often the executive best positioned to ensure that AI transformation happens with and through the workforce.
A well-built AI hub supports that transformation. It turns fragmented pilots into value at scale. It connects technology deployment with workforce adoption. It embeds responsible governance into delivery. And it gives the CEO and executive team a clearer line of sight from AI investment to business impact.
The authors thank Savannah Ricard and Frank Anderson for their contributions to this publication.
