Context-aware AI becomes the battleground for the next technology race

AI needs 'context.' Enterprise AI is delivering impressive results in recommendations, automation, and rapid decision support, but experts argue that without understanding how a company actually operates, even the most powerful models can easily miss the right call. After years of experimentation and pilot projects, executives are now demanding clear evidence that scaling up AI actually improves business performance. AI agents operating without operational context, the explanation goes, can run into limits by applying general principles directly to specialized situations.

Debabrata Shah, chief scientist for enterprise AI at Celonis, sees this context as central to the concept of 'decision intelligence.' AI agents can support individual tasks within a larger decision-making process, he noted, but they must understand what role each task plays and the surrounding business conditions. "They can't do it in a vacuum," he said.

Without the right operational context, agents risk overlooking critical constraints, making costly mistakes, or burning excessive computing resources searching for information they need. A context layer enables agents to understand workflows, operate within established guardrails, and produce recommendations that humans can review and approve.

Shah advises starting with a single business decision, identifying its most difficult steps, and testing whether AI can improve them. He saw opportunities expanding into company-wide planning areas such as finance, workforce management, marketing, and sales. With the right context in place, AI can help evaluate multiple options and make critical decisions faster. "You no longer need to spend so much time on resource planning, and people can make the right choices quickly," he said.

This contributed article was published by MIT Technology Review Insights on October 22, 2026, under the title 'Making the case for context-aware AI,' through a partnership with Celonis, a process mining (a technology that analyzes a company's actual workflows through data) company.