BRIAN KRZANICH
APPLICATIONS
He notes:“ Enterprise environments introduce constraints that most consumer applications never encounter: noise, latency, trust and security, worker safety, regulatory requirements, limited connectivity and highly specialised operational workflows.”
As organisations deploy AI across their operations, the interaction layer becomes increasingly important.“ People need a reliable and intuitive way to access information, direct intelligent systems, approve actions and understand outcomes,” Brian points out.
BRIAN KRZANICH
TITLE: CEO COMPANY: CERENCE AI INDUSTRY: AUTOMOTIVE LOCATION: MASSACHUSETTS, US
Brian serves as President and CEO of Cerence AI, taking the reins in 2024. He was previously CEO of Intel, having joined in 1982 as a Process Engineer and rising through the ranks.
A critical capability“ Conversational AI is a critical capability, but understanding language alone is not enough in physical environments,” Brian asserts.
He highlights that, in those more demanding environments, systems need to understand what’ s happening around them:“ Machines, equipment and industrial processes constantly generate signals about their condition and performance, and interpreting those signals can provide important context that isn’ t available from language alone.”
This is an area where Cerence has developed deep expertise through decades of commercial deployments in automotive.
“ By bringing together conversational AI and audio AI,” Brian goes on,“ intelligent systems gain a richer understanding of both human intent and operating conditions.
“ That combination becomes especially valuable in manufacturing, robotics and industrial environments, where context often matters as much as the command itself.”
Brian states that the most effective AI systems increasingly rely on specialised models working together, coordinated through an orchestration layer that manages context, determines which capabilities should be applied to a given request and ensures a consistent user experience.
CaLLM Brian explains that, with CaLLM, Cerence AI’ s proprietary family of automotive-grade large and small language models, it has realised domain-specific optimisation truly matters.
“ Models tuned for the realities of a vehicle, a manufacturing environment or a particular workflow consistently outperform general-purpose approaches in those settings,” he says.
86 October 2026