AI Leadership:
SHAPE-AI

Why Some Organizations Scale AI While Others Struggle

While organizations continue to invest heavily in artificial intelligence, many struggle to translate experimentation into measurable business results. SHAPE-AI identifies the leadership behaviors associated with accelerating AI adoption, building trust, and turning AI investments into organizational impact.

A 2025 MIT report cited in Harvard Business Review found that many generative AI initiatives had not yet produced measurable bottom-line returns. ghSMART’s research examines a different question: the leadership behaviors associated with turning AI investment into adoption and business impact.

When ghSMART asked 53 senior leaders to rank the single biggest driver of AI return on investment, 47% put leadership effectiveness first – far ahead of workflow integration (15%), organizational culture (11%), and engineering talent (8%). In this group, leadership effectiveness ranked well ahead of the other factors offered.

At the same time, only 32% of respondents to ghSMART’s AI Leadership Velocity Survey said their organizations had the right leaders in place to succeed with AI. The finding points to a leadership challenge as well as a technical one: organizations need leaders who can translate AI investment into changes in workflows, decisions, and performance.

Successful AI organizations need two kinds of leaders: AI Architects, who build the technical foundation, and AI Shapers, who drive adoption, trust, and business impact. Technical capability alone is not enough; organizations also need leaders who can translate AI investment into organizational change and measurable results.

What Is SHAPE-AI?

A Behavioral Framework for AI Leadership, Not a Technical Assessment

SHAPE-AI is not a measure of technical AI fluency. It is a framework of five behavioral dimensions associated with effective AI adoption and impact: Strategic Agility, Human Centricity, Applied Curiosity, Performance Drive, and Ethical Stewardship. These behaviors may not be visible on a resume and are distinct from technical AI expertise.

The framework is designed to assess AI Velocity. AI Velocity is ghSMART’s proprietary concept for how quickly and effectively an organization moves from AI experimentation to scaled, measurable results. AI Velocity considers the leadership conditions that support innovation, trust, adoption, and change. SHAPE-AI is used to assess the behaviors that may accelerate or constrain that progress.

The central premise of SHAPE-AI is that effective AI leadership requires more than technical fluency. It also requires specific behaviors that support learning, adoption, execution, and responsible use.

Research Basis – Interviews, Survey Data, and HBR Publication

SHAPE-AI draws on primary research and ghSMART’s broader leadership-assessment work. It draws on interviews with more than four dozen senior executives across Fortune 50 corporations, private-equity portfolio companies, and nonprofits; ghSMART’s AI Leadership Velocity Survey, which captured proprietary data across industries; and a literature review of 38 relevant studies.

The findings were published in Harvard Business Review in September 2025 by Rens van den Broek, Samantha Hellauer, and Dina Wang, and in NACD Directorship that same month by Rens van den Broek and Samantha Hellauer – bringing the framework to both the executive and the boardroom audience.

SHAPE-AI applies ghSMART’s behavioral assessment approach to the specific demands of AI leadership. It builds on more than 30 years of leadership research and the same behavioral research tradition as ghSMART’s CEO Genome research and the Who Method. SHAPE-AI extends ghSMART’s broader leadership research portfolio, including CEO Genome®, the Who Method, and the Potential Model, into the emerging domain of AI leadership.

Two Complementary Archetypes – Architects and Shapers

SHAPE-AI distinguishes between two leadership archetypes that successful AI organizations need in combination.

AI Architects are the technical leaders who design and govern the foundational AI infrastructure. They ensure the data architecture, tooling, and governance are in place for scalable, responsible AI use. Many sit in the CTO or CIO organization. Their work makes enterprise AI possible.

AI Shapers are the business leaders who embed AI into workflows to drive performance, adoption, and innovation. They do not need deep technical expertise. They need curiosity, influence, and judgment to translate AI’s potential into results. Most AI initiatives are led by non-technical functions, which means the Shaper role extends well beyond IT into finance, marketing, operations, HR, and the front line.

A small group of technical experts cannot provide all of the leadership required to embed AI across an organization. Shaper capabilities need to extend beyond the technology function and into senior business leadership. The CEO must model curiosity. The CFO must reimagine financial processes. The CHRO must rethink talent decisions. SHAPE capabilities cannot be concentrated in a single AI champion or a newly minted Chief AI Officer; they must be distributed across teams, functions, and levels of the enterprise.

Both archetypes apply the SHAPE framework – but the five behaviors manifest differently depending on whether a leader is building the system or driving its adoption. The sections that follow show that distinction dimension by dimension.

The Five SHAPE-AI Dimensions

S – Strategic Agility: Navigate Fast Without Losing Direction

Strategic Agility is the ability to navigate fast-moving environments while staying focused on long-term business goals. In AI, leaders must respond to new models, new capabilities, new competitors. Great leaders balance speed with judgment – knowing when to accelerate, when to pause, and when to let go of what no longer creates value. What it looks like in practice:

• Prioritize options over rigid plans, and proactively scan for disruption before it derails progress.
• Focus on business value rather than novelty, and avoid the sunk-cost trap of defending an initiative because it was expensive to start.
• Make real-time adjustments and continuously align AI efforts to ROI.

For Architects: rapidly pivot the technical stack as model capabilities evolve.

For Shapers: quickly adjust rollout plans when adoption lags or new use cases emerge.

Failure mode: making linear plans that ignore uncertainty, defending work because of prior investment, and selecting tools without connecting them to a broader goal.

Survey finding: Strategic Agility was ranked the most important SHAPE dimension – 65% of respondents placed it first or second – and simultaneously the hardest to find and develop. Because respondents viewed Strategic Agility as both important and difficult to develop, it may warrant particular attention in selection and development.

H – Human Centricity: Trust Sets the Speed Limit for AI Adoption

Human Centricity is the ability to drive AI adoption by preserving trust, fostering inclusion, and anchoring transformation in shared purpose and values. Trust can materially affect the pace of AI adoption, particularly when employees are being asked to change established ways of working. What it looks like in practice:

• Don’t frame the question as humans versus AI; ask instead, “How can AI make our people better?”
• Design change with people, not for them, and model AI use personally and visibly.
• Build psychological safety through empathy and feedback loops.
• Address employee fears actively rather than assuming adoption will take care of itself.

For Architects: design systems with explainability and usability built in – human-in-the-loop by default.

For Shapers: model responsible use and cultivate a culture of experimentation and inclusion.

Failure mode: rolling out change top-down with little input, framing AI primarily as a cost-cutting efficiency play rather than a way to elevate human contribution, and blaming resistance on others instead of revisiting the approach.

Survey finding: Human Centricity ranked among the hardest SHAPE capabilities to find – yet it was also seen as the most coachable. The finding suggests Human Centricity may be particularly responsive to deliberate development.

A – Applied Curiosity: Learn, Test, and Apply

Applied Curiosity is the ability to explore emerging technologies with both imagination and discernment – constantly learning, questioning, and applying insights to drive meaningful business outcomes. In AI, curiosity is most useful when it is paired with disciplined experimentation and learning. What it looks like in practice:

• Combine systematic scanning with disciplined experimentation, running fast, cost-effective tests with clear learning objectives.
• Filter hype by asking a single question: “Is this solving our problem, or someone else’s?”
• Build readiness through purposeful experiments rather than chasing shiny objects.
• Learn from the front lines – junior employees and field teams often surface the most valuable AI use cases.

For Architects: continuously test new models and platforms while separating signal from noise.

For Shapers: scan for emerging capabilities that unlock value and encourage experimentation at the team level.

Failure mode: delegating exploration to colleagues instead of role-modeling it, experimenting without ever drawing conclusions, and chasing headlines without testing fundamentals.

Survey finding: Applied Curiosity ranked second most important – 47% placed it first or second – yet only about one-third of organizations feel strong in it. It was also rated the second hardest dimension to coach, which makes early identification critical: this capability may deserve greater weight in hiring as well as development.

P – Performance Drive: Connect AI to Business Results

Performance Drive is the ability to turn AI potential into performance – channeling energy and resources into real business results. In AI, that means maintaining discipline around business outcomes, ROI, and scaling. What it looks like in practice:

• Reject activity metrics in favor of business outcomes, and sunset low-impact initiatives rather than keeping them on life support.
• Establish execution rhythms with clear accountability and cross-functional alignment.
• Maintain momentum through continuous, measurable impact.

For Architects: build scalable, cost-effective systems and optimize model costs and data pipelines.

For Shapers: identify high-leverage cross-functional use cases and drive coordinated change across teams, systems, and silos.

Failure mode: celebrating pilots that never scale, measuring success with activity metrics instead of business outcomes, and letting novelty outshine value.

Survey finding: Performance Drive is seen as more coachable over time than Strategic Agility or Applied Curiosity. But the gap it addresses is wide – only about half of respondents say their organization consistently connects AI initiatives to enterprise performance. This gap between experimentation and sustained business value is central to the challenge of scaling AI.

E – Ethical Stewardship: Build Responsibility Into AI from the Start

Ethical Stewardship is the ability to lead AI with integrity – embedding responsibility into the design, deployment, and governance of technology from the start. Effective AI leaders build responsibility into design, deployment, and governance rather than treating it as a later-stage requirement. What it looks like in practice:

• Build governance that manages risk while accelerating progress – not governance as an afterthought.
• Make transparency and human oversight standard operating procedure.
• Treat algorithmic bias as a business risk equivalent to financial or operational risk.

For Architects: embed ethics, bias mitigation, and governance into the core architecture, and engage externally to help shape emerging standards.

For Shapers: integrate responsible-AI practices into business workflows and ensure equitable access to AI tools across teams.

Failure mode: treating governance as an afterthought, downplaying bias risks until they become public problems, and prioritizing launch speed over responsible implementation.

Survey finding: Ethical Stewardship consistently ranked last on priority lists – yet interviews showed it becomes critical the moment organizations move beyond pilots into scaled deployment. The finding suggests that governance can become more important as organizations move from pilots to scaled deployment.

The SHAPE-AI Evidence: What the Data Shows

Organizations Are Weakest Where They Need Strength Most

The survey showed that leaders rated their organizations as weakest in the very SHAPE dimensions they identified as most important. The capability seen as most decisive – Strategic Agility – was also named the hardest to find and coach, creating a leadership-development challenge. The widest gaps between stated importance and actual organizational strength appeared in Ethical Stewardship and Human Centricity: the capabilities organizations lack most are precisely the ones they perceive as hardest to develop. Left unaddressed, that mismatch does not close on its own – it widens with every scaled deployment.

The Architects vs. Shapers Imbalance

Most organizations have invested in AI Architects – the technical leaders who build the systems. The shortfall is in AI Shapers: only one in three organizations believe they have leaders who can play that role today.

The imbalance cuts both ways. When Architects build technically robust systems without Shapers to drive adoption, those systems may fail to take hold, limiting the value of the technical investment. Shapers who surface valuable use cases likewise depend on Architects to deliver reliable, scalable solutions. The framework treats the two archetypes as complementary.

The Board Dimension – From Oversight to Acceleration

The leadership gap does not stop at the C-suite; it extends into the boardroom. While 55% of board directors say their boards discuss AI, 41% do so only once or twice a year. Fewer than a quarter are satisfied with the board time devoted to the topic, and only 21% of leaders believe their boards understand AI well enough to govern it effectively.

The board does not need to become a group of AI experts. Its role is to ensure the right executives are in place, set the tone for experimentation and accountability, and anchor AI aspirations in long-term organizational values. Boards can support AI performance by setting expectations, overseeing leadership capability, and connecting AI investment to strategy and risk.

How SHAPE-AI Is Used in Practice at ghSMART

Assessment – Mapping the Leadership Bench

ghSMART uses SHAPE-AI to assess leaders not only for individual roles but for the strength of an organization’s overall AI leadership bench. The assessment looks past technical fluency to the behavioral dimensions – Strategic Agility, Human Centricity, Applied Curiosity, Performance Drive, and Ethical Stewardship – that drive or stall AI Velocity.

The output is a SHAPE-AI profile that maps strengths and gaps by dimension across both Architects and Shapers. It informs CEO and C-suite succession, board-level talent reviews, and pre-deal leadership due diligence in private-equity contexts. The framework builds directly on ghSMART’s assessment methodology – see the Who Method – and connects to the Assess & Select Leaders hub for organizations evaluating their bench against AI demands. For the engagement view of this work, see the SHAPE-AI solution page.

Development – Building the Muscle That Matters

SHAPE capabilities are buildable – but not equally coachable. Strategic Agility and Applied Curiosity are harder to develop quickly; Human Centricity and Performance Drive respond more readily to targeted coaching. Effective development relies on stretch roles, coaching grounded in real business context, tight feedback loops, and safe-to-fail experimentation environments.

The approach scales. In one Fortune 50 application, ghSMART assessed the top 300 next-generation leaders against SHAPE-AI to map the leadership bench, diagnose strengths and weaknesses, and deliver tailored development across the population. The Potential Model – ghSMART’s CQ/DQ/EQ framework – can provide complementary insight into the underlying behaviors and growth capacity relevant to AI Shapers.

Hiring – Acquiring Strength Where It’s Hardest to Build

Because Strategic Agility and Applied Curiosity are the hardest SHAPE dimensions to develop, they are the most important to hire for. Prior AI experience is rare and should not be the primary criterion – it may exclude leaders who have the underlying behaviors needed to learn and lead effectively. The focus shifts to learning agility, tolerance for ambiguity, intellectual curiosity, and the ability to work across silos.

SHAPE-AI gives hiring teams a behavioral framework for evaluating AI leadership alongside role-specific requirements. Applied through the Who Methodology and the SmartAssessment Scorecard, it sharpens executive selection and management-team assessment for roles where AI fluency will be learned but leadership behavior must already be present.

Board and Investor Use Cases

Boards use SHAPE-AI to assess whether the C-suite has the AI leadership capability to execute on the strategic commitments the board has approved. In private equity, the framework supports pre-deal and post-deal assessment of the leadership bench for AI readiness, and ongoing evaluation of portfolio-company leadership.

For boards, the practical priorities are to select and support the right leaders, shape the conditions for AI leadership to flourish, and anchor AI aspirations in organizational values.

Conclusion

The choice facing leaders is not only whether to invest in AI, but whether to build the leadership required to turn that investment into results. Organizations that scale AI successfully pair technical capability with leaders who can drive adoption, learning, trust, and performance. SHAPE-AI provides a framework for identifying, assessing, and developing those leadership behaviors.