Creating AI value with enterprise context

Prukalpa Sankar, co-founder of Atlan, frames the shortfall as a capability-versus-usefulness gap. "There's a big gap between what the models are capable of and how the models are actually being useful inside organizations," he says. "For every real-life intent and purpose, artificial general intelligence is already here. It's just not really useful in most companies today." Sankar argues the missing layer is context: an architectural layer sitting between an organization's business systems — systems of record and data — and its AI platforms, translating enterprise intelligence into governed, machine-usable context that agents and models can consume and that lets knowledge compound. Without it, he likens the situation to hiring the smartest intern in the world, blindfolding them, and throwing them into your organization.

The poll numbers and Sankar's diagnosis together imply that the bottleneck has shifted from model quality to organizational plumbing: whether a company can convert its systems of record into governed, machine-readable context that agents can act on. For builders and operators, the practical question is what moves the 12% figure — how many teams can take agents from pilot into production on top of a context layer rather than bolting them directly onto raw data.