How to stop AI costs from soaring

According to research from analyst Gartner, 30% of generative artificial intelligence (GenAI) projects are expected to be abandoned after the proof-of-concept phase by the end of 2025. This is primarily due to challenges such as poor data quality, rising costs, and a lack of clear business value.

Gartner’s findings show that early adopters in various industries and business processes have experienced different levels of business improvements based on the use case, job type, and worker skill level. On average, respondents reported a 15.8% increase in revenue, 15.2% cost savings, and 22.6% productivity improvement.

When it comes to deriving value from AI, Eyad Tachwali, a senior director at Gartner, emphasizes the importance of understanding the different ways AI can be utilized. He distinguishes between everyday AI, which enhances existing tasks, and game-changing AI, which focuses on creativity and innovation.

While everyday AI aims to boost productivity, game-changing AI is geared towards generating new ideas and solutions. For example, a pharmaceutical company using AI to discover a new molecule for drug development showcases the potential of game-changing AI.

When implementing GenAI applications, IT leaders must consider various factors to determine the necessary investments. Factors such as use cases, industry, and organizational risk appetite play a crucial role in cost estimation. While off-the-shelf products like ChatGPT offer straightforward cost calculations based on user numbers and software licenses, game-changing AI initiatives involve additional costs such as IT infrastructure, data, and application development.

Tachwali highlights the importance of understanding the multiplicative cost elements associated with generative AI models. Token-based pricing, for instance, can significantly increase costs by five to ten times. Gartner recommends simulating usage limits of large language models to better manage costs and ensure that the benefits outweigh the expenses.