Why Enterprise AI Struggles to Deliver Transformative Value, According to New Insights Data

Many organizations deploying multiple AI tools find themselves struggling to extract transformative returns from their investments. According to a poll by MIT Technology Review Insights, while every surveyed organization currently uses artificial intelligence to some degree, only a small fraction report achieving transformative outcomes from their deployments.

The findings highlight a widespread friction point in enterprise technology: advanced models and platforms running simultaneously across customer platforms, conversational assistants, and automation tools often produce outcomes that lack a proper understanding of the specific business.

Understanding the Return on Investment Gap

The MIT Technology Review Insights survey data reveals that fewer than one in five organizations—specifically 16%—are extracting transformative value from their AI deployments. Meanwhile, the largest segment of respondents, 40%, report seeing moderate gains limited to specific tasks or teams. Another 21% report meaningful gains, while 23% observe minimal returns or no return at all.

In terms of deployment maturity, companies remain heavily focused on early stages. The poll indicates that 38% of respondents are still piloting AI use cases, and 35% are primarily using AI to assist employee workflows through tools like co-pilots and chat assistants. By comparison, only 12% of enterprises have deployed AI agents into actual production environments, and just 3% report utilizing AI to handle the majority of work in key business processes.

The Need for an Enterprise Context Layer

Prukalpa Sankar, co-founder of Atlan, points to a disconnect between the raw capabilities of modern models and their practical utility inside corporate structures. According to Sankar, artificial general intelligence effectively exists for many real-life intents, but it remains largely unhelpful within typical companies due to a missing component: business context.

This missing layer serves as an architectural infrastructure bridging an organization's core business systems—such as systems of record and underlying data—and its deployed AI platforms. Its function is to translate enterprise intelligence into governed, machine-usable context that AI models and agents can reference, allowing institutional knowledge to compound over time. Without this structural layer, implementing advanced AI models resembles onboarding an exceptionally capable worker while depriving them of necessary organizational awareness.

Practical Implications and Verification

For technology leaders and enterprise strategists, these findings underscore that deploying sophisticated AI models alone is insufficient without accompanying architectural frameworks that connect models to internal data systems.

• What is confirmed: Survey statistics regarding enterprise AI adoption stages, return levels reported by respondents, and expert commentary concerning the role of architectural context layers.
• What remains unknown: The exact methodology, sample size, and specific industry breakdown of the poll respondents beyond the high-level percentage figures published in the custom content report.
• How to verify: Readers can review the original publication details and full custom insights report directly through the [MIT Technology Review official source](https://www.technologyreview.com/2026/10/28/1147709/creating-ai-value-with-enterprise-context/).