AI Investment Falters: 95% of Companies See No Return on Generative AI
Table of Contents
Despite widespread enthusiasm, a new study reveals that the vast majority of organizations are failing to realize a measurable return on their generative AI investments. The report, published by MIT in July, serves as a stark reminder that simply adopting the technology doesn’t guarantee success.
The findings indicate that 95% of organizations are not seeing quantifiable benefits from their AI spending, highlighting a critical gap between hype and reality. Successfully scaling AI projects beyond initial pilot phases is now paramount to transforming the current wave of excitement into genuine return on investment (ROI).
the Pitfalls of Broad Implementation
One key challenge lies in the overly enterprising approach many companies are taking. According to a senior official at NTT Data, an IT services and consulting company, a common mistake is attempting to integrate AI into every aspect of operations. “Companies say, ‘In every single domain, I’m going to unleash innovation, and I’m going to have AI enablement.’ I think thatS the wrong strategy,” the official stated at the Fortune Global forum in Riyadh on sunday.
Instead, a more focused approach is recommended: identifying one or two areas where AI can deliver important economic value and pursuing complete implementation within those domains.Examples cited include focusing on underwriting in insurance and optimizing supply chains in manufacturing.
Prioritization and Strategic Integration
Companies that demonstrate a clear AI strategy, prioritizing specific applications, are far more likely to succeed. FedEx, for instance, is intentionally integrating AI into three core areas: internal operations, customer experience, and the creation of new customer value thru improvements in areas like demand forecasting and reducing returns.
“Research has shown that organizations that have a clear AI strategy,which has this prioritization,have a much greater degree of success than others that don’t,” explained a FedEx president overseeing operations in the Middle East,Indian Subcontinent,and Africa.”For us, that’s the key aspect of scaling.”
The Human Element and the Risk of “Hallucinations”
Deploying AI also necessitates robust safeguards and ongoing human oversight. A CEO of Vortexa, a cargo tracking and energy market research firm, emphasized the importance of explainability – the ability to understand why an AI model is making a particular decision. “In addition to speed and quality, what is…increasingly vital on any model, any agent, is explainability. For a human to be able to understand, why is it making the decision that it is making?”
This is particularly crucial in sensitive fields like healthcare, where the potential for AI “hallucinations” – confidently presenting fabricated details – poses a significant risk. the founder and CEO of January AI, a precision healthcare company, noted that while large language models (LLMs) have access to vast amounts of data, they are prone to inventing information when faced with gaps in their knowledge. “These LLMs, yes, they have read everything, but they do hallucinate, and they do it confidently. When there is data missing, they will invent it,” she said, adding that her company’s Mirror tool achieves a hallucination rate of under 1% through rigorous doctor review.
Data Silos Hamper Healthcare AI Adoption
The healthcare industry faces unique obstacles to scaling AI,particularly in the United States. Fragmented data, spread across patients, insurers, providers, and laboratories, hinders the growth of a unified view necessary for effective AI implementation.
“Not having a unified view of data doesn’t really allow us to leverage AI in the best way that we can,” one healthcare executive explained. Despite the technological capability to potentially eradicate lifestyle-based chronic diseases, progress is hampered by data silos, regulatory hurdles, and privacy concerns.The question, she posited, is whether there is sufficient collective will to overcome these challenges and fully deploy the transformative potential of AI.
