Latest News & Resources

The decline in AI performance, highlighted by recent reports of ChatGPT becoming less effective, should prompt caution among those eager to adopt AI without thorough readiness assessments, feasibility studies, or AI transformation strategies. The rush to implement AI without proper preparation can lead to significant challenges and setbacks.

Noise and Redundancy in AI Discussions

The current discourse around AI is filled with repetitive information. In the past month alone, I’ve encountered numerous executive reports and analyses that essentially echo each other. It wouldn’t be surprising if some of this content was generated by ChatGPT itself. Here, I aim to provide a more strategic approach for those who believe in planning before acting.

Reframing the Terminology

We should shift from calling it ‘artificial intelligence’ to ‘augmented intelligence.’ This change in terminology can bring clarity and set a more accurate perspective for enterprises planning to implement AI. When businesses view this technology as an enhancement to their workforce, they will focus on the right question: How do we supercharge our workforce with AI?

Strategic Technology Adoption

Adopting technology is not a race. It’s about creating a sustainable long-term strategic advantage. While generic AI use cases might be served by SaaS providers, the real impact will come from enterprise-wide applications and differentiation strategies, such as private LLMs with Retrieval Augmented Generation (RAG) fine-tuned to specific data.

Enhancing Workforce Productivity

Integrating AI as a co-pilot can significantly improve communication quality and productivity across an organization, especially for non-native English speakers. However, there are risks. Over-reliance on AI for creating, editing, and paraphrasing content can lead to uniformity and a lack of originality. For technical tasks, such as code generation, the risk is higher if developers do not fully understand the AI-generated code.

The Burden of AI Integration

Each SaaS and ERP provider will promote their AI capabilities, but the responsibility of integrating and understanding these tools will fall heavily on organizations. In my book on Cloud computing, I argued that as enterprises adopt various cloud solutions, the effort to consolidate and make sense of their data increases due to fragmentation across systems. This challenge extends to AI, where predictions and data can become siloed.

Strategic Planning for AI Adoption

Use 2023 and early 2024 to experiment and learn about AI, LLMs, base models, fine-tuned models, and other tools. Simultaneously, start defining your data strategy and identify use cases that provide strategic advantages. By Q2 of 2024, focus on selecting a tech stack and developing use case POCs that can scale to enterprise-wide solutions.

Human Oversight in AI Decision-Making

Before relying on AI for decision-making, consider the case study of Volvo, which ended up with more green cars than they started with due to the lack of human oversight. Always ensure a human-in-the-loop to prevent such issues.

By taking these steps, organizations can harness the potential of AI while mitigating risks and ensuring sustainable growth.