Photo by Kevin Ku on Pexels

Generative AI’s initial buzz is fading, and businesses are now scrutinizing its real-world ROI. Concerns are mounting, with an MIT NANDA report finding that a staggering 95% of AI pilot projects fail to scale or deliver measurable returns. Some experts even suggest that relentlessly chasing ROI metrics is misguided. However, tangible AI success is achievable by shifting focus to three core principles. First, treat data as a valuable asset. Instead of simply preparing data for AI models, explore leveraging its exclusivity to negotiate favorable terms with model vendors, trading access for services or discounts. Second, prioritize reliability over novelty in AI deployments. Focus on implementing robust, stable AI solutions for routine, essential tasks – ‘boring by design’ – rather than constantly pursuing the latest and greatest models. Finally, embrace ‘mini-van economics,’ designing AI systems with a focus on user needs and cost-effectiveness. Streamline workflows to minimize reliance on expensive third-party services. By prioritizing practicality, fostering independence, and capitalizing on their unique data assets, companies can unlock the true potential and demonstrable ROI of their AI investments.