The rise of generative AI (gen AI) has been propelled by prominent large language models (LLMs) like OpenAI's GPT-4, Google's Gemini, and Anthropic's Claude. However, smaller language models (SLMs) are emerging as potential frontrunners in the field, according to some experts. Research from Gartner indicates that although LLMs have led language model development, SLMs may address critical challenges for businesses such as budget constraints, data protection, privacy, and risk mitigation. Businesses might find themselves weighing the benefits of larger versus smaller models in their exploration of gen AI. Five business leaders share their insights on this topic: 1. **Domain-Specific Models**: Claire Thompson, from L&G, envisions a future where both LLMs and SLMs coexist, especially as LLMs are fine-tuned for specific topics. She believes in developing tailored models for domains like healthcare and finance but is skeptical about companies committing resources to build them from scratch. 2. **Choosing the Right Model**: Nick Woods of MAG Airports Group advocates for a hybrid approach, selecting models based on specific business use cases rather than adopting a one-size-fits-all strategy.

He emphasizes focusing on business transformation and suggests deploying small models to address particular needs. 3. **Contextual Application**: Gabriela Vogel from Gartner highlights the growing trend of companies utilizing small, context-specific models tailored to particular applications, as organizations move towards implementing more focused gen AI solutions. 4. **Minimizing Hallucinations**: Ollie Wildeman from Big Bus Tours notes that smaller models often provide better results for businesses, as they are designed with specific data in mind. This approach enhances data safety and reduces inaccuracies, leading companies to prefer domain-specific models. 5. **Utilizing First-Party Data**: Rahul Todkar from Tripadvisor emphasizes that the effectiveness of AI models lies in their ability to be customized to fit the business's context and utilize first-party data effectively. He argues that the future of AI will revolve around creating tailored models rather than simply selecting between large and small. In summary, while LLMs attract attention, the discussion suggests that SLMs may hold significant promise, particularly for specific applications and contexts in business.