Intelligence, the startup behind the AI evaluation platform Design Arena, has raised $7.9 million in seed funding to expand its human-feedback system for improving artificial intelligence models.
The funding round was led by Index Ventures, with participation from Conviction, A*, Valkyrie, and other investors. The investment will support the company’s mission to provide scalable human evaluation for AI-generated content.
Founded in 2025 by Grace Li and a group of college classmates, the company originally set out to build an AI-powered game engine. While the technology could generate functional games, the team discovered that determining whether a game was genuinely enjoyable still required human judgment.
That realization led to the creation of Design Arena, a platform that allows users to compare AI-generated outputs through simple “A vs. B” rankings. The system helps AI developers collect real-world human preferences to improve the quality of their models.
According to the company, Design Arena is now used by 5.3 million people worldwide and has become an important source of evaluation data for leading AI developers. Intelligence says the platform currently generates $60 million in annual recurring revenue (ARR) through enterprise customers that rely on its feedback system.
Unlike traditional AI benchmarks, Design Arena captures user preferences across different formats, including websites, images, and other visual content. Because users log in before accessing results, the company can also analyze how design preferences vary by region and evolve over time.
The startup believes human evaluation remains essential for training AI systems, particularly as automated benchmarks become increasingly vulnerable to manipulation. Recent incidents, including security issues affecting AI benchmarking platforms, have highlighted the importance of reliable human-led assessment.
Despite growing investor interest in AI evaluation startups, the market remains competitive. Some companies, such as Yupp, failed to establish sustainable business models despite attracting significant funding. However, others, including LM Arena, have secured major investments as demand for human feedback in AI development continues to grow.
