Harvard's Raffaella Sadun on why it's so hard to become an AI-first organization

Key Concepts

Frontier Firm: A company that actively learns and strategically experiments with new technologies like AI to drive business value and empower its workforce.
J Curve of Productivity: A concept describing the initial dip in productivity when implementing new technology, followed by a subsequent rise. In the context of AI, each firm is on its own unique J curve.
Tacit Knowledge: Knowledge that is difficult to articulate or transfer, often held by experts. AI has the potential to make this knowledge more accessible.
Reskilling: The process of teaching employees new skills to adapt to technological changes and evolving job requirements.
Paradigm: A model or framework. The discussion outlines five paradigms for reskilling.
Strategic Experimentation: A process of trying out new approaches and technologies within defined strategic boundaries to understand their potential impact and value.
Human-Machine Collaboration: The integration of human capabilities with AI agents to enhance productivity and decision-making.
Agent Employee: The concept of an AI acting as a digital assistant or employee, requiring management and guidance.
Scientific Method for Learning: Applying hypothesis testing, measurement, and learning to understand what works and what doesn't in technological adoption and reskilling.
Peer Learning: The process of learning from colleagues and other organizations through shared experiences and insights.

Frontier Firms and Strategic Experimentation

Raffaella Sadun, Professor of Business Administration at Harvard Business School, discusses the concept of a "Frontier Firm" in collaboration with Microsoft. A Frontier Firm is defined as an organization that prioritizes learning and invests in strategic experimentation to understand how new technologies, particularly AI, can benefit them. This involves taking learnings from experimentation back into the organization.

Patterns for Successful Reinvention:

Clarity of Vision and Value Creation: Firms need a clear understanding of their core purpose, how they create value, and what differentiates them. This clarity guides experimentation within specific strategic boundaries, such as improving products, services, or reducing costs.
Organizational Buy-in and Adoption: Once the value of AI is understood, it's crucial to bring the entire organization along. This involves addressing fear of replacement and ensuring new ideas are adopted beyond a small group of enthusiasts.
No Universal Playbook: Sadun emphasizes that there is no single playbook for implementing AI. Companies must accept this uncertainty and develop strategies that fit their unique culture and people, integrating technology, strategy, and organizational structure.

Human-Agent Collaboration

A foundational element of a Frontier Firm is the concept of every employee potentially managing an "agent employee" (AI). This represents a significant mindset shift.

Managing an Agent Employee:

Managerial Skills are Key: Just as not everyone is a good human manager, managing AI agents requires specific skills. This includes asking the right questions, understanding what can be delegated, and critically evaluating the AI's responses.
AI as an Eager Apprentice: AI agents are described as incredibly knowledgeable and having vast memory, but they don't always get it right, may struggle to admit ignorance, and require specificity and context. Onboarding and training these agents is crucial.
Improving Human Management: The skills developed for managing AI agents can also lead to better management of human employees, addressing a known area for improvement in many organizations.

Executive Time Allocation and Leadership Styles

Research into how executives allocate their time reveals that leadership approaches need to adapt to the evolving technological landscape.

Phased Leadership Approach:

Early Stage (Figuring out Technology): Leaders need to be close to the action, potentially redesigning value creation processes. This proximity builds credibility and ensures resources are allocated to what's truly important.
Adoption and Diffusion Stage: Leaders must identify key players within the organization to build buy-in. This involves understanding the mindset of skeptics and enthusiasts and leveraging internal "heroes" or ambassadors to influence others.
Metrics and Compliance: While important, Sadun notes that creating metrics focused on compliance too early can hinder the exploration phase. Metrics should initially support exploration and the creation of new processes.

Reskilling in the Age of AI

Reskilling is presented as a critical but challenging aspect of adapting to new technologies, as it requires individuals to fundamentally change their identities and roles.

Five Paradigms for Reskilling:

Business Rationale and Clarity: Clearly articulate why reskilling is necessary and what the future holds for employees who undergo it. This provides certainty and a sense of future opportunity.
Shared Responsibility: Reskilling should not be solely an HR initiative but a shared responsibility across the entire organization, especially among those reporting to the CEO.
Middle Management Engagement: Middle managers play a crucial role. Some are natural coaches, while others may resist. Building their cooperation and understanding is vital.
Employee Perspective: Companies must consider the employee's viewpoint, clearly outlining the promise of reskilling in terms of career progression, incentives, and future direction. Employees need to see the benefit beyond the immediate cost.
Ecosystem and Supply Chain Collaboration: Leverage knowledge within the supply chain and broader ecosystem to reskill at scale, potentially creating shared learning initiatives. This is termed "It Takes a Village."

Techniques for Effective Reskilling:

Beyond Passive Training: Reskilling is more than just sitting through passive training sessions. It requires a fundamental shift in identity and role.
Combination of Hard and Soft Skills: Effective reskilling programs integrate both technical (hard) skills and interpersonal (soft) skills.
Community and Cohort Learning: Learning within a community or cohort, rather than in isolation, leads to better outcomes and engagement.
Reskilling as a Journey, Not an Episode: Reskilling should be an ongoing process integrated into the firm's strategic goals, not a one-off event.

Strategic Imperative and Investment:

Strategic Imperative: Companies must view reskilling as a strategic imperative, not just a training expense.
Investment Levels: Companies are already investing significant portions of their budgets (e.g., 1.5% of total budgets, as per a BCG report) in training. The key is to invest well.
Designing for Learning and Practice: Resources should be spent on programs that genuinely help people learn and practice new skills, acknowledging that adults are not always intrinsically motivated to learn.
Scientific Mindset for Investment: Implement reskilling investments with a scientific mindset, including clear hypotheses, measurement, and a focus on learning what works and what doesn't.

Transitioning from Pilot to Sustainable Change

Moving from experimentation to sustainable process change requires a recognition that it's more than just adopting new technology.

Key Steps for Sustainable Change:

Leadership Mandate and Clarity: A clear leadership mandate and understanding of why change is happening are essential.
Learning from Others: Adopt a wide lens to observe how other organizations are adopting technologies. Go to where the action happens and learn from networks and peers.
Strategic Boundaries for Experimentation: Experimentation should occur within defined strategic boundaries (e.g., 1-3 key strategic axes) to maintain coherence.
Cross-Functional Experimentation Teams: Assemble teams with technical expertise, an understanding of organizational use, and representation from potential resistors to identify and mitigate frictions.
Integrating Adoption from the Start: Build the adoption and scaling phase into the early stages of experimentation, ensuring new initiatives are perceived as extensions of existing work, not foreign objects.
Reverse Engineering Metrics: Reverse engineer metrics and measurement by considering the desired end state of massive adoption and how people and organizations evolve.

Actionable Advice

For CEOs:

Strategy Session: Conduct a focused strategy session to define 2-3 key objectives for experimentation.
Experimentation Team: Carefully select an experimentation team comprising technical experts, human-centric individuals, and those who might represent potential resistance.
Communication Plan: Develop a clear communication strategy to explain what is being done, what is being learned, and what it means for the organization.

For Employees:

Recognize Your Role as an Inventor: Embrace the mindset of an inventor if you want to make an impact.
Develop Technical Skills: Acquire necessary technical skills.
Ensure Fit with Organizational Goals: Understand how your work aligns with the organization's broader objectives.
Build a Movement/Coalition: Connect with other enthusiasts to share learnings and aggregate ideas, fostering collective growth.

The Future of Work (3-5 Years)

Sadun anticipates massive divergence across firms in how jobs are performed, even within the same occupation. This divergence is a consequence of firm-specific investments and evolving tasks.

Implications for Employees and Hiring:

Job Evolution: Employees should expect their jobs to look significantly different in 3-5 years, especially in firms undergoing rapid change. The match between an employee's aspirations and the firm's direction is crucial.
Hiring for Adaptability: Hiring will increasingly focus on an individual's ability to learn new skills rather than solely on existing experience.

Conclusion

The discussion highlights that navigating the AI era requires a strategic, human-centric approach. Frontier Firms are those that embrace continuous learning, strategic experimentation, and foster human-agent collaboration. Reskilling is a complex but vital process that demands clarity, shared responsibility, and a deep understanding of employee needs. The future of work will be characterized by significant divergence, emphasizing the importance of adaptability and a strong organizational fit for both employees and employers.