iKala Discusses Agentic Commerce, GEO Trends, and Updates TMMLU+ Benchmark
Frontier AI models · AI assistants · Enterprise AI
iKala notes that AI agents are expanding from search queries into purchasing decisions, shifting marketing from SEO to Generative Engine Optimization (GEO). The team also highlighted TMMLU+ v1.1, an open-source benchmark evaluating model comprehension of Traditional Chinese and Taiwan local context.
According to reports, iKala Co-founder and Chairman Sega Cheng and AI Marketing Director Chiao-Tung Hsu discussed shifts in AI commerce dynamics. Cheng stated that AI agents are advancing from query responses to task execution, enabling Agentic Commerce where AI may make purchasing decisions on behalf of consumers. Under this trend, marketing is moving from traditional SEO to Generative Engine Optimization (GEO), requiring brands to manage visibility and citations in AI responses. For internal enterprise adoption, iKala proposed a strategy of rapid validation in key scenarios followed by gradual expansion. To address cultural and linguistic context, iKala maintains the open-source TMMLU+ benchmark, releasing version v1.1 to evaluate comprehension of Traditional Chinese and Taiwan knowledge.
Key takeaways
- Sega Cheng stated that AI agents are driving Agentic Commerce, shifting purchasing decisions and transactions to AI execution.
- Brand marketing is moving from SEO to GEO, requiring focus on how AI models describe brands, cite sources, and recommend products.
- Addressing enterprise AI adoption, Cheng noted leaders must understand AI, with iKala advising rapid validation in key scenarios.
- To test model understanding of Traditional Chinese and Taiwan context, iKala maintains the open-source TMMLU+ benchmark and released v1.1.
Source：數位時代 Read the original ↗
Prepared, checked and automatically published by AI; not individually reviewed by a human. Translation checked by AI.