Most companies have yet to move beyond early artificial intelligence experiments: 38% of respondents are piloting AI use cases, while 35% use copilots or chat assistants to support employees. Just 12% have AI agents in production, and only 3% use AI to carry out most of their key processes. The gap, according to Atlan co-founder Prukalpa Sankar, is not the capability of AI models but the business context they receive.
AI adoption remains concentrated in pilot projects
Companies are primarily testing AI rather than deploying it across core operations. In the survey, 38% of respondents said their organizations were piloting AI use cases. Another 35% said they were using AI to assist employees through tools such as copilots and chat assistants.
Only 12% of companies reported having AI agents in production. Just 3% said they use AI to perform most of their key processes.
“There is a huge disconnect between the power of a model and how it can actually help an organization,” says Prukalpa Sankar, co-founder of Atlan. “For all real-world intents and purposes, artificial general intelligence already exists. It’s not really useful in most companies today.”
Why business context matters
Sankar says the missing component is context: how well an AI system understands a particular business. This context layer is the architectural infrastructure between an organization’s business systems—including systems of record and data systems—and its AI platforms.
It converts enterprise intelligence into context that managed machines, AI agents and models can use to build on existing knowledge. “Without the context layer of AI, it’s like hiring the smartest intern in the world and throwing them blindfolded into your organization,” Sankar says.
This content was created by Insights, the custom content division of MIT Technology Review, and not by editorial staff. It was researched and written by humans, and any AI tools that may have been used were limited to production processes under human supervision.
Source: www.technologyreview.com