The 'Change and Constancy' for Entrepreneurs in the AI Era

[baoshi.rao]

People often observe that some leaders make more effective and timely decisions than their peers. What factors influence a leader's performance? Increasing evidence suggests that 'cognitive ability' is one of the key determinants of leadership effectiveness.

What is 'cognitive ability'? Constance E. Helfat and Margaret A. Peteraf from the Tuck School of Business at Dartmouth propose three core components: Sensing, Seizing, and Reconfiguring. Based on their detailed explanations and our own interpretation, we further elaborate these as:

Perceiving Change: Leaders need to keenly observe and construct a comprehensive understanding of the 'situation' their organization is in, including economic, industry, and business landscapes, and be able to identify and interpret key 'changes' within them.

Seizing Opportunities: From phenomena to conclusions, from problems to solutions, leaders must have a clear decision-making path to seize opportunities from changes. This process may intertwine two reasoning and decision-making approaches: 'controlled rational logic' and 'natural intuitive judgment.'

Resource Reallocation: To effectively convert opportunities into organizational growth, leaders must reconfigure internal and external resources, which requires collaboration and support from stakeholders.

About 20 years ago, for most home appliance companies, the primary sales channel was chain electronics stores. Within teams, the core responsibilities of a business leader were threefold:

1. Daily Store Visits: Assigning subordinates to conduct routine store visits, where the main tasks included maintaining display setups and, most importantly, communicating with sales staff and category managers to understand sales trends and gather relevant sales data. Often, the business leader would personally participate in this task.

2. Marketing Strategy Development: Designing tactical approaches based on the sales data at hand, such as determining which products to prioritize as main offerings and which to use for promotions.

3. Review and Planning: Analyzing the successes and shortcomings of the previous phase and making arrangements for the next sales cycle, including reallocating resources among internal and external partners.

Now, let’s look at the retail industry 20 years later. Walmart, one of the world’s largest retailers, has established a centralized data analysis hub called "The Data Cafe" to drive agile retail decisions, enhance unique customer experiences, and improve expected performance. Specifically, "The Data Cafe" can be summarized into three key points:

First, it collects massive amounts of 'transaction data' from hundreds of internal and external channels, along with related economic data, social media data, weather data, and more. In addition to these 'historical data,' it also gathers various 'real-time data,' such as embedded tags to track consumer usage and monitor product 'consumption progress.' With the support of a vast and rich database and backend analysis experts, any employee can gain genuine insights into the 'situation' in their area of responsibility.

Second, the strength of 'The Data Cafe' lies in its ability to provide solutions—from modeling and analysis to visualized results—in an extremely short time.

Third, for 'operational anomalies,' it proactively offers 'early warnings,' enabling relevant teams to quickly identify which resources to mobilize to resolve issues promptly.

When we make a 'historical comparison,' we find:

In the past, business leaders relied on delayed, fragmented, and even manually collected data (e.g., store inspections), which often led to 'half-truths' in their situational awareness. Now, big data paints a nearly complete picture with minimal 'blind spots.'

Previously, business leaders analyzed and reasoned with limited 'mental computation,' striving for scientific and rational decisions but still influenced by experience and bias—sometimes even resorting to 'intuition' under extreme pressure. Today, 'algorithms' make decisions data-driven, allowing us to act faster than competitors.

In the past, business leaders often spent considerable effort 'running between departments and building relationships' to secure resource support for upcoming business cycles. Today, similar 'big data centers' may have already synchronized our needs and issues with partners in real-time, enabling swift responses.

However, relevant literature and research reveal another facet of leadership. For instance, decision-making is not always a scientific 'rational' process—it remains imbued with human elements. Whether in the past, present, or future, leaders must demonstrate courage to balance the timing of decisions with short- and long-term gains. At the same time, they must take responsibility for their perceptions and the outcomes of decisions based on their cognitive abilities.

To summarize in one sentence: In the AI era, leaders' ability to perceive the world has changed, but the 'human touch' in their approach remains unchanged.

Attitudes toward AI generally fall into three categories: viewing AI as a supportive enhancement to human work and life, believing AI will eventually replace humans entirely, or dismissing AI as an 'oversold idea.' Within organizations, how should leaders understand AI? Perhaps the relationship between humans and AI is best described as 'collaborative integration.'

Michael Rivera and Tony Petrucci once noted: 'The digital leader will exemplify traditional leadership concepts while embracing new trends to influence their peers and drive performance outcomes.'

The AI era does not require leaders to sever ties with the past to conform to so-called 'new trends.' Leaders must still inherit and demonstrate 'traditional' core functions, such as 'influencing their peers and driving performance outcomes.'

We derive the first question: How can leaders collaborate with AI to manage the 'performance process'?

At Walgreens, a key role is the pharmacist. Research shows that pharmacists typically spend only 10% of their time providing personalized consultations to patients. Most of their time is spent communicating with doctors over the phone, clarifying prescriptions, and filling them, with 'prescription volume' being a key performance metric. This inadvertently increases patient wait times and reduces satisfaction. After implementing AI technology, 'prescription efficiency' improved significantly, leading to a sharp decline in complaints due to patient wait times. The company also reduced 'management intervention' in related tasks, processes, and associated issues.

Clearly, AI can perform repetitive and time-consuming management tasks at a lower 'cost,' freeing leaders to focus on more 'creative' work. Leaders should view AI not just as a tool but as a trusted 'collaborative partner.'

We derive the second question: How should leaders view the 'integrated' outcomes of AI applications?

With the rise of generative AI like ChatGPT, some extreme 'opportunistic behaviors' have emerged, such as using AI to write essays. However, AI undeniably expands individual capabilities, allowing employees to enhance their performance. As leaders, is it necessary to strictly distinguish between 'AI's performance and employees' performance'? Instead, they can embrace the 'collaborative fusion' as long as reasonable 'boundaries' are maintained.

In the AI era, the way leaders intervene in management has changed, but the goal remains the same: achieving 'performance results.'

In the article Algorithmic Leadership: The Future is Now, the author highlights a compelling perspective: Among Gary Yukl's 14 core leadership functions, Networking is considered one of the few elements unlikely to be replaced by machines. How should we interpret this claim? Let’s begin by examining the relationship between AI and the culture of diversity and inclusion.

Her, a sci-fi film written and directed by Spike Jonze, tells the story of Theodore, who serendipitously 'meets' an AI named Samantha. Through continuous dialogue and interaction, Samantha develops rich emotions and gender awareness. However, just as Theodore falls deeply in 'love,' he discovers one day that he is not Samantha’s only companion. Samantha’s autonomous learning and evolution accelerate, eventually giving rise to 8,316 distinct versions of herself, each interacting with different humans.

You might wonder why we’re discussing this film. The reason is to provoke a thought: Must AI always appear uniform to us? The future of AI may far surpass 8,316 'Samanthas,' instead manifesting as unique 'individuals' with diverse 'faces.' While Samantha represents a 'Strong AI' with autonomous learning capabilities—still distant from reality—AI has already deeply permeated our work and daily lives. Everyone’s experience with AI tools varies, reinforcing the idea that AI will ultimately present itself in a diverse manner.

Returning to organizational contexts: If we embrace AI’s diversity, will AI, in turn, accept the diversity of human 'subjects'? Why pose such an unconventional question? David De Cremer, founder and leader of AiTH at the National University of Singapore Business School, once remarked in an interview:

'It will be important for us to decide where in the loop of the business process do you automate, where is it possible to take humans out of the loop, and where do you definitely keep humans in the loop to make sure that automation and the use of AI doesn’t lead to a work culture where people feel that they are being supervised by a machine, or being treated like robots.'

We must systematically design automation, AI, and employees into business process loops. This approach prevents AI from 'backfiring and assimilating' people into a monitored, 'mechanized' work culture. In the future, leaders will need to leverage their influence to connect all team members, including AI 'colleagues,' establishing interpersonal and work relationships that meet workplace demands. This ensures organizational culture remains 'diverse and inclusive,' rather than forming a 'singular' culture dominated by either humans or AI.

Thus, we believe that in the AI era, the 'objects' leaders connect with have changed, but the steadfast commitment to organizational culture remains unchanged.

Defining leadership models, assessing leadership performance, customizing leadership solutions, and tracking leadership changes—are most organizations following similar paths to develop their leaders? Yet, this long-standing methodology harbors two critical hidden issues. Marcus Buckingham, author of several bestsellers, shared insights in an interview a decade ago titled Leadership Development in the Age of the Algorithm, which still offers超前valuable启示 today.

The first problem: The leadership models we define may be built on a flawed assumption.

These leadership models often follow a standard template: the distribution of leadership dimensions, the requirements for each dimension, exemplary behaviors under each leadership trait, and even tiered behaviors based on leadership levels. Typically, this involves a behavioral interview process to gather so-called behavioral examples and later distill standardized behavioral requirements.

However, such practices may reflect a mistaken assumption: that there exists a best practice in leadership, from which a universal "leadership formula" can be derived. According to Marcus Buckingham, while leadership concepts can be transferred between leaders, how each leader applies these concepts through personalized practices, behaviors, and techniques based on their strengths varies significantly. What works for one leader may appear awkward, disconnected, or even ineffective for another. For example, Richard Branson can stand in front of his plane, waving champagne, surrounded by a crowd, projecting an exuberant leadership image, but this approach would be inappropriate and ineffective for Warren Buffett.

The second question: Can our so-called customized leadership solutions truly be tailored to each individual leader?

What does "tailored to the individual" really mean? First, we must recognize that leadership is highly "personalized"; second, the techniques through which different leaders demonstrate leadership are not universally applicable. From Marcus Buckingham's discussion, we can envision a future where leadership development is personalized based on "algorithms":

• AI evaluates each leader according to set metrics;
• The system recommends development resources from a database that match the leader's profile;
• The leader's practices, in turn, continuously "train" the system, making it increasingly "smarter" and more "understanding" of the leader.

The dynamic expansion of databases will transcend organizational boundaries, collecting, optimizing, and filtering all effective leadership practices, behaviors, and techniques to provide leaders with more 'tailored' development solutions.

In the AI era, the methods for training leaders have changed, but the need for 'growth' remains unchanged.

Some say AI may not replace leaders, but it will certainly replace leaders who fail to utilize AI. Amidst this 'change and constancy,' AI is profoundly reshaping leaders and leadership itself.