An Eye-Opening Report

Earlier today, I stumbled upon an insightful blog post by Sheryl Estrada, a senior writer at Fortune, and I just had to share it with you all.

Sheryl dives deep into the current state of generative AI in business, and her findings are both eye-opening and thought-provoking. In her piece, Sheryl highlights a striking reality: while companies are eagerly investing in AI, a staggering ninety-five percent of generative AI pilot programs are failing to deliver meaningful results.

This information comes from a report titled "The GenAI Divide: State of AI in Business 2025," published by MIT's initiative.

The report reveals that although generative AI has immense potential, most initiatives aimed at driving rapid revenue growth are falling flat.

In fact, only about five percent of these AI pilots are achieving the kind of rapid revenue acceleration that businesses are hoping for. Sheryl's research is based on extensive interviews with leaders, a survey of three hundred fifty employees, and an analysis of three hundred public AI deployments.

The findings paint a clear picture of a divide between those companies that are successfully leveraging AI and those that are struggling to make it work.

Success Stories and Their Secrets

To unpack these insights, Sheryl spoke with Aditya Challapally, the lead author of the report.

He pointed out that while some large companies and younger startups are excelling with generative AI, the majority of companies are facing significant challenges. For instance, startups led by young entrepreneurs have seen their revenues soar from zero to twenty million dollars in just a year.

The secret? They focus on one specific pain point, execute their strategies effectively, and forge smart partnerships with companies that utilize their tools.

This targeted approach is a game changer. However, for the vast majority of companies in the dataset, the implementation of generative AI is falling short.

The Real Challenge: Learning Gaps

The core issue isn't the quality of the AI models themselves but rather a "learning gap" that exists within both the tools and the organizations using them.

While executives often point fingers at regulations or model performance, Sheryl's findings suggest that the real problem lies in flawed enterprise integration.

Tools like ChatGPT may excel for individual users due to their flexibility, but they often stall in enterprise environments because they don't adapt to existing workflows. Another critical insight from Sheryl's article is the misalignment in resource allocation.

Where ROI Truly Lies

More than half of the budgets for generative AI are being funneled into sales and marketing tools. Yet, the research indicates that the highest return on investment is actually found in back-office automation.

This includes eliminating business process outsourcing, reducing costs associated with external agencies, and streamlining operations. So, what factors contribute to successful AI deployments? Sheryl emphasizes that how companies adopt AI is crucial.

Purchasing AI tools from specialized vendors and building partnerships leads to success about sixty-seven percent of the time.

In contrast, internal builds succeed only about one-third of the time.

This is particularly relevant in highly regulated sectors like financial services, where many firms are attempting to create their own proprietary generative AI systems.

Partnerships Over Going Alone

However, the data suggests that companies experience far more failures when they go it alone. Another key takeaway is the importance of empowering line managers—not just central AI labs—to drive adoption.

Selecting tools that can integrate deeply and adapt over time is also vital for success.

Workforce Shifts and Shadow AI

As we look ahead, workforce disruption is already underway, especially in customer support and administrative roles.

Instead of mass layoffs, companies are increasingly choosing not to backfill positions as they become vacant.

Most of these changes are concentrated in roles that were previously outsourced due to their perceived low value. Sheryl also touches on the widespread use of "shadow AI," which refers to unsanctioned tools like ChatGPT.

This presents an ongoing challenge when it comes to measuring AI's impact on productivity and profit.

The Next Phase of Enterprise AI

Looking forward, the most advanced organizations are already experimenting with agentic AI systems—those that can learn, remember, and act independently within set boundaries.

This offers a glimpse into how the next phase of enterprise AI might unfold. In conclusion, Sheryl Estrada's blog post is a must-read for anyone interested in the future of AI in business.

It's a powerful reminder that while the potential of generative AI is vast, the path to successful implementation requires thoughtful strategy, smart partnerships, and a willingness to adapt.

So, let's take these insights to heart as we navigate the exciting world of technology and storytelling together!