Why it matters
Front-end experimentation driven by agents changes how A/B testing and personalization loops operate. Engineers integrating agents into experimentation platforms need to understand the latency and call-volume tradeoffs.
The tokenmaxxing angle
Agentic experimentation compounds token spend: each variant evaluation may trigger multiple LLM calls. Understanding how to batch, cache, or constrain agent calls within feature-flag pipelines keeps per-experiment costs predictable.
From the organizers
Event description states the evening is focused on 'AI agents and experimentation' with speakers Benjamin Odisho and Mike Bosch presenting 'Optimizing Agents for Front-End Experimentation' at Merchandise Mart.