How CEOs Can Turn Generative AI Into a Superpower: A 5‑Step Framework

McKinsey’s latest report, Reconfiguring work: Change management in the age of gen AI , reveals that while two‑thirds of global firms have adopted generative AI, its impact remains vague. The study argues that generative AI is not merely a tool but a "super‑ability" that can reshape work, and CEOs must adopt new change‑management strategies to move employees from passive experimenters to active accelerators.

Step 1 – Define an Outcome‑Driven North Star Vision

Leaders should treat generative AI as a capability, not a mandatory tool, and craft a simple, inspiring North‑Star vision that explains how AI will create value and competitive advantage while considering talent lifecycle effects. The vision must be flexible enough to absorb rapid AI model evolution, allowing organizations to start with single‑task AI agents and evolve toward "agent swarms" that deliver end‑to‑end business outcomes. Some departments may become Minimal Viable Organizations (MVOs) led by AI agents with minimal human supervision, while others retain more human involvement for high‑touch customer work.

Step 2 – Build Trust Through Data Access, Governance, and AI Oversight

Trust is a prerequisite for scaling AI. High‑performing firms that attribute at least 10% of EBITDA to generative AI invest heavily in trust‑building activities such as risk mitigation and governance, which double the likelihood of revenue growth above 10% compared with peers.

Organizations should treat data access as a first‑class workflow, enabling AI to consume unstructured data while addressing challenges of training models on mixed data types. CEOs, CIOs, and CDOs should establish an AI oversight committee, define acceptable‑use policies, and embed "human‑in‑the‑loop" checkpoints to guard against hallucinations, bias, or data leakage—especially critical in regulated sectors like finance.

Example: Morgan Stanley trained an internal assistant on over 100,000 research reports using OpenAI, deployed it only after rigorous quality evaluation, and achieved a 98% adoption rate among wealth‑management teams, dramatically democratizing expert knowledge.

Step 3 – Reshape Workflows and Evolve Toward AI‑Centric Teams

Simply adding AI to existing processes yields only incremental gains. The report proposes a "two‑in‑one" approach where business and technology teams co‑design end‑to‑end workflows that balance business outcomes with technical feasibility.

The evolution occurs in three phases:

Phase 1: Deploy independent AI agents to handle discrete tasks (e.g., specific steps in procure‑to‑pay), where AI functions as a tool.

Phase 2: Form agent swarms that collaborate across entire processes under human supervision.

Phase 3: Achieve fully autonomous agent swarms (MVOs) with humans only overseeing high‑level outcomes, freeing staff for higher‑value work.

Early employee involvement is crucial; selecting clear‑value, low‑risk processes for pilot projects boosts buy‑in. Training reduces anxiety—48% of U.S. employees say they would use AI more with training, and 45% would use it more if integrated into daily workflows.

Step 4 – Hybrid MVOs and Enhanced Teams for Organizational Redesign

After AI integration, CEOs must decide which functions become streamlined MVOs (e.g., invoice processing with AI‑driven matching and approval) and which retain human‑centric roles (e.g., sales, customer service) that benefit from AI‑augmented capabilities. MVOs require AI‑ops monitoring and new talent such as AI workflow optimizers; many existing staff will need reskilling.

While back‑office functions can achieve large cost reductions, front‑office areas still need human touch to preserve brand experience.

Step 5 – Empower Employees as AI Ambassadors and Continuous Learners

Transformation requires organization‑wide participation. Studies show that companies involving more than 7% of employees in AI initiatives double the probability of delivering above‑average total shareholder return, with high‑performers achieving 21‑30% excess returns.

Employees should be treated as AI ambassadors; “super‑users” (e.g., millennial managers) can drive cultural change. CEOs must model AI usage, while middle managers cascade the practice.

McKinsey’s internal platform Lilli exemplifies this approach: 92% of global staff have used it, 74% use it regularly, saving over 30% of time and generating 19 million AI‑generated suggestions. The platform hosts clubs, one‑on‑one executive courses, and a marketplace for employee‑built agents (1.7 × 10⁴ additional agents) without adding technical debt.

Singtel’s AI Acceleration Academy, launched in October 2024, trained 10 000 employees across roles to use generative AI, fostering cross‑functional innovation.

Conclusion – Co‑Creating the Future of Work

CEOs who begin planning a company‑wide redesign today will enable humans and AI to co‑create extraordinary outcomes tomorrow. Change management must nurture an experimental culture, turning employees from passive users into proactive change agents. Generative AI is not a threat but a hidden, indispensable colleague that can free staff from repetitive tasks and amplify creativity, reshaping competitive dynamics.