Key findings, 2023 to 2024 To illustrate the evolving landscape, we Do you customize generative AI models or use them out of the box? see the following changes as important Organizations reporting trends in AI over the past year. generative AI forced the creation of an AI strategy: Organizations with no Organizations planning plans to work with to increase investment in generative AI: commercial and closed- source models over the next three years: Organizations that Organizations with consider AI to be very generative AI models in or highly critical to production: their business in the next three years: Organizations applying AI are seeking to extract specific tasks or datasets, significantly enhancing their additional value by optimizing AI through prompt-en- performance and accuracy on targeted applications. 2023 2024 gineering, fine-tuning models, and retrieval augmented RAG further enhances this by dynamically incorporat- 2024 generation (RAG). Despite the desire to optimize foun- ing external information during the generation process, dational models, 65% of organizations use models out- enabling the model to produce more relevant and con- of-the-box, 43% of organizations fine-tune models and textually rich outputs. 38% use RAG. Fine-tuning can customize models for 06 07