Integrating Generative AI Into Business Strategy: Dr. George Westerman

The 2025 MIT Bangkok Symposium featured Dr. George Westerman, a Senior Lecturer at MIT Sloan School of Management. He discussed how organizations can integrate generative AI into their business strategies. The session focused on understanding AI's rapid evolution, its impact on various business tasks, and the organizational capabilities needed to succeed with this technology. Dr. Westerman emphasized making informed decisions about AI adoption, regardless of how fast technology changes.

Understanding AI: More Than Just Technology

Dr. George Westerman, a leading expert in digital transformation, shared his insights on integrating generative AI into business strategy. He highlighted that while technology changes quickly, organizations often change much slower. This means the real challenge isn't adopting new tech, but transforming how a business operates.

Dr. Westerman introduced "Westerman's Law": Technology changes quickly. Organizations change much more slowly. This idea is key because it means the hard part of AI adoption is not the technology itself, but changing business processes and leadership approaches. He stressed that technology provides zero value on its own; value comes from how a company uses it to change its business or products.

He also pointed out a crucial fact: Artificial intelligence is not intelligent. It's a program that executes based on what it's been programmed to do or learned. It doesn't have context knowledge. As Aude Oliva, an MIT expert, puts it, "artificial intelligence should be artificial idiots." However, it can act intelligently, which is useful if used carefully.

Opportunities for Digital Transformation with AI

AI is the next stage of digital transformation, offering powerful opportunities. Dr. Westerman outlined four key areas where companies can look for opportunities, not just with AI, but with other technologies like mobile and IoT:

• Emotionally Engaging Customer Experience: Creating personalized and targeted experiences for customers.

• Operations: Focusing on adaptability and adjustment, beyond just automation (Industry 4.0).

• Business Models: Using information to create new services or improve existing ones, like turning products into services.

• Employee Experience: Recognizing that satisfied employees lead to satisfied customers, and a poor employee experience often signals systemic issues.

These areas are where AI can make a real difference, especially when combined across different functions, as seen with companies like Home Depot and Airbus.

Generative AI in Action: Real-World Examples

Generative AI is already being used in many practical ways:

• Content Creation: Generating course content, translating corporate literature into multiple languages instantly.

• Coding Assistance: Tools like Copilot help programmers write code, enforce standards, and create documentation more efficiently.

• Call Center Support: Cresta, a tool for sales call centers, provides real-time hints to agents, improving performance for both senior and junior staff.

• Personalized Learning: Developing AI tutors for educational programs, helping students overcome challenges and pursue new career paths.

• Integration into Existing Products: Major software providers like SAP, Workday, and Adobe are integrating generative AI into their platforms.

Dr. Westerman emphasized that the best solutions often combine generative AI with traditional AI, existing IT systems, and human processes. For example, Lemonade, an insurance company, uses a mix of AI and traditional systems to automate 50% of claims, leaving complex cases for human review. Sysco, a food service delivery company, uses AI across its operations, from customer experience to warehouse routing, showing how even non-tech companies can benefit.

Categorizing AI: A Practical Approach

To make sense of the rapidly evolving AI landscape, Dr. Westerman proposed four categories of AI, offering a practical way to understand their applications and limitations:

Rule-Based Systems (Expert Systems)

• Description: Based on "if/then" statements, programmed by talking to experts.

• Pros: Provides precise and consistent answers; no data needed.

• Cons: Does not adapt; limited to simple problems; difficult to add new rules.

Econometrics (Statistics)

• Description: Uses statistical methods on structured, often numeric, data.

• Pros: Works well with numeric data; relatively cheap to program; provides precise and consistent answers.

• Cons: Requires numeric data; can have false positives/negatives.

Deep Learning

• Description: Uses neural networks trained with labeled data to make predictions.

• Pros: Can handle complex patterns; outputs are repeatable.

• Cons: Requires large amounts of labeled data; outputs are not easily explainable; susceptible to data bias.

Generative AI

• Description: Creates new content by predicting the next best word or element based on vast amounts of data.

• Pros: Creates new things, not just classifies; can be used for creative tasks.

• Cons: Can produce "hallucinations" (incorrect or fabricated information); requires huge training data and energy; answers are random and not always the same.

Making AI Work in Your Organization

Implementing AI successfully requires addressing organizational challenges, not just technical ones. Dr. Westerman highlighted three main challenges:

• Prioritization: Deciding what AI initiatives to pursue first, second, and never.

• Risk Management: Addressing concerns like privacy and potential errors.

• Capabilities: Building the necessary skills and processes within the organization.

Key Takeaways:

• Start with the problem, not the technology. Define the business problem first, then choose the right AI technique.

• Consider accuracy and cost of error. How critical is accuracy? What are the consequences of being wrong?

• Evaluate explainability. Do you need to understand why the AI made a certain decision?

• Assess data availability and quality. Do you have reliable, unbiased data to train the AI?

• Address confidentiality. Ensure data privacy and security.

Organizational Readiness and Culture

Successfully integrating AI also depends on an organization's culture and its people. Dr. Westerman emphasized:

• Cultural Readiness: Is the culture open to working with AI, even when it challenges human expertise? Companies need humility to embrace AI and an ethical framework to use it responsibly.

• Experimentation: Fostering a culture that encourages experimentation and learning from failures.

• Skills and Careers: AI will change jobs, but it doesn't have to replace people. Instead, it can make jobs easier, reduce cognitive load, and serve as a powerful teaching tool. Organizations should focus on helping employees adapt and grow with AI, rather than fearing job displacement.

The Path to AI Transformation: Small Steps, Big Impact

Many companies are not pursuing massive, immediate AI transformations. Instead, they are focusing on "transformation with a little t" – smaller, systematic changes that build capability over time. This approach involves:

1. Individual Productivity: Using AI for tasks like summarizing documents or updating spreadsheets, often with low risk.

2. Specialized Roles and Tasks: Applying AI to specific functions like call centers or coding, often with a human still in the loop.

3. Direct Customer Impact: Engaging customers directly through AI, such as personalized online shopping experiences or first-tier customer service.

This incremental approach, often called the "risk slope," allows companies to grow their risk management capabilities alongside their AI adoption. It's like changing a car tire: small, balanced adjustments lead to a stable outcome. This systematic learning helps organizations move from lab-based proofs of concept to real-world implementation, addressing errors and continuously improving along the way.

Conclusion

Dr. Westerman's presentation offered practical advice for navigating the world of generative AI:

• Be intelligent in how you use AI. Understand its limitations and remember that imperfections don't make it useless; implement controls just as you would for human error.

• Start with the problem, not the technology. Focus on business needs, and be open to combining different AI techniques.

• Get started now. Begin with small, manageable projects to build experience and capabilities.

• Prepare your people. Help employees understand how AI can enhance their jobs and involve them in the adoption process.

• Continuously improve. Use small transformations to pave the way for larger, more impactful changes over time.

By following these principles, organizations can effectively integrate generative AI, driving innovation and achieving lasting value.