Creating AI value with enterprise context

Enterprise AI initiatives often utilize a multitude of platforms, systems, and tools concurrently. These include AI-driven customer solutions, conversational agents, and workflow automation technologies. While each component is advanced and responsive, the outcomes they generate may be inaccurate, misguided, or insufficiently contextual to the business.

Numerous organizations encounter this challenge, which may explain their struggle to realize returns on their AI investments. According to an MIT Technology Review Insights survey, while every surveyed organization employs AI to some extent, fewer than one in five (16%) are deriving transformative value from their deployments. The majority (40%) report moderate gains in specific tasks or teams, 21% indicate meaningful gains, and 23% observe minimal returns or no returns at all.

Most companies (38%) are still in the pilot phase for AI use cases or leveraging AI as an assistant to employees' work through co-pilots or chat tools. Only 12% of enterprises have AI agents deployed in production, and just 3% claim to utilize AI for the majority of work in critical processes. According to Prukalpa Sankar, co-founder of Atlan, "There's a significant gap between what the models are capable of and how the models are actually being useful inside organizations."

Sankar asserts that the missing component is a context layer—a technical framework that bridges an organization's business systems (such as systems of record and data) and AI platforms. This layer translates enterprise knowledge into structured, governable context that AI agents or models can utilize, allowing knowledge to accumulate over time.

Sankar likens the absence of a context layer to hiring the most intelligent intern but blindfolding them and thrusting them into the organization without understanding its workings.