Babak Hodjat

Hodjat has been arguing for the primacy of agent-based AI architectures since the late 1990s, when the field was still a niche academic research area. In this keynote, he brings that long view to bear on the current moment, explaining why agentic AI and multi-agent systems represent not just the latest trend but the natural architectural endpoint of what AI has always been trying to become. He covers the specific advantages of multi-agent approaches over centralized LLMs for complex enterprise tasks, the orchestration and governance challenges that determine whether agentic deployments succeed or fail, and the concrete applications his team at Cognizant is delivering at scale. A technically grounded, forward-looking keynote for organizations trying to understand what comes after the current generation of generative AI tools.

Long before large language models dominated the AI conversation, Hodjat was building AI systems that used evolutionary computation, genetic algorithms, and distributed processing to solve problems that rule-based and gradient-descent approaches could not. In this keynote, he explains the evolutionary AI paradigm, its specific advantages for optimization, strategy discovery, and model architecture search, and why it is experiencing a renaissance as organizations look for AI approaches that are more robust, interpretable, and capable of handling genuinely novel problems. He draws on his work at Sentient Technologies, his research at Cognizant's Advanced AI Lab, and his 39 US patents to present a view of AI that goes well beyond the current LLM mainstream.

Hodjat's career traces a through-line in AI that few other speakers can offer: from the agent-oriented natural language architectures he invented at Dejima in the late 1990s, through the distributed evolutionary AI systems he built at Sentient, to the multi-agent enterprise architectures he is deploying at Cognizant today. In this keynote, he uses that arc to give audiences a structural understanding of where AI has come from, why the current moment is genuinely different from previous inflection points, and what the patterns of the past three decades suggest about where the technology is heading. One of the most historically grounded and technically credible AI perspective keynotes available.

Hodjat leads AI deployment for a company with over 350,000 employees and clients across every major industry sector. In this keynote, he draws on that operational reality to address the question most AI strategy sessions avoid: what does it actually take to move AI from successful pilots to sustained production systems that change how an enterprise operates? He covers the architectural decisions that determine scalability, the governance and evaluation frameworks that keep enterprise AI systems reliable, the organizational and cultural barriers that derail deployments far more often than technical ones, and the specific patterns his team has found most effective in translating AI investment into measurable business outcomes.