麦肯锡-充分发挥生成式人工智能在银行业的价值（英）-9页_422kb

Summary of Generative AI in Banking

Key Benefits and Opportunities

Generative AI (Gen AI) is poised to transform banking by adding up to $340 billion annually to operating profits through increased productivity and innovation. It can enhance customer engagement, content synthesis (virtual experts), content generation, and software development, with initial applications focusing on efficiency gains. However, scaling these pilots to capture full value is challenging due to factors like rapid change, talent gaps, and evolving operating models.

Challenges in Scaling Gen AI

Scaling Gen AI is distinct from traditional AI due to:

• The broad scope requiring rethinking roles and strategies.
• Disrupting established operating dynamics between business and technology.
• The unprecedented pace of adoption, stressing existing models.
• Unique talent challenges, including upskilling and recruitment.

Seven Dimensions for Successful Scaling

1. Strategic Road Map: Requires senior leadership alignment, clear goals, priority domains, and enabling capabilities planning.
2. Talent: Leaders must upskill or attract expertise in prompt engineering, model fine-tuning, and AI integration; address concerns about automation and job roles.
3. Operating Model: Avoid centralized misnomers; promote cross-functional teams for seamless implementation and continuous feedback.
4. Technology: Balance build vs. buy decisions, integrate Gen AI with legacy systems, and maintain a coherent architecture.
5. Data: Leverage unstructured data through Gen AI tools; ensure data quality, security, and infrastructure alignment.
6. Risk and Controls: Manage risks like bias, intellectual property, privacy, and reliability by embedding controls and governance from the start.
7. Adoption and Change Management: Focus on user-centric design, training, and incentives to ensure full adoption and avoid resistance.

Risks of Gen AI

Gen AI introduces unique risks, including algorithmic bias, IP issues, privacy concerns, security threats, lack of explainability, reliability issues, organizational disruption, and ESG impacts. Mitigation involves expert validation, automation, and transparent frameworks.

Scaling Gen AI effectively requires a holistic approach addressing these dimensions to unlock sustained value while managing risks.