Logan Kilpatrick, Senior Product Manager at Google DeepMind, has called on AI companies to develop their own benchmarks to better assess AI model performance. Speaking on X, Kilpatrick emphasized that custom benchmarks allow companies to focus on metrics relevant to their specific business needs, rather than relying on public leaderboards that may not reflect their unique use cases. He highlighted that companies like Zapier and Sierra are already benefiting from this approach, which can drive significant improvements in AI model performance tailored to business-specific tasks.