What Lies Behind Successful AI Deployment?

MIT's report suggests that how companies adopt AI is crucial. The success rate of purchasing AI tools from specialized providers and establishing partnerships is about 67%, while building internally has only one-third of that success rate (about 22%).

This finding is particularly relevant in financial services and other highly regulated industries, where many companies are building their proprietary generative AI systems by 2025. However, MIT's research indicates that these go-it-alone companies face a much higher failure rate.

Challapally noted that surveyed companies are often reluctant to share failure rates. "Almost everywhere we go, companies are trying to build their own tools," he said, but the data shows that purchased solutions deliver more reliable results.

The report also reveals that AI-driven disruption in the labor market has already begun, particularly in customer support and administrative roles. Unlike large-scale layoffs, companies are increasingly choosing not to refill vacant positions. Most changes are concentrated in jobs previously outsourced due to perceived lower value.

Arguably, the release of this report comes at a time when concerns about high valuations of tech stocks are growing. The market has once again become highly sensitive to any negative news related to artificial intelligence—as evidenced by the turmoil in the U.S. stock market triggered by the emergence of DeepSeek in January.

Tuesday's sell-off further indicates that after months of AI-driven frenzy, any evidence questioning the commercial viability of AI could trigger another market correction.

In terms of trends, Tuesday's decline was driven by the drop in some of this year's best-performing stocks. Oracle and Advanced Micro Devices (AMD), two of the top five large-cap stocks since mid-May, fell by 5.9% and 5.4%, respectively. AppLovin, which places ads within apps, dropped by 5.9%.