Kando AI raises tens of millions of yuan in seed round to become 'Cursor for decision-making'
The AMW Read
Novelty is incremental: Kando is the latest in a wave of vertical AI workbenches targeting high-value decision domains, but the 'Cursor-for-decision' framing is a clear analog of a known pattern. Significance is sub-segment-level: the seed round is small, the product is early, and the outcome is unc
Kando AI raises tens of millions of yuan in seed round to become 'Cursor for decision-making'
Kando AI, a Beijing-based startup founded in June 2026 by Beida CS PhD Wu Bingzhe and serial entrepreneur Mao Shuhan, has raised tens of millions of yuan (~$1.4M-$4.2M) in a seed round led by Xinglian Capital with participation from Lihui Gold Holdings and an unnamed industrial investor. The company is building an AI workbench that it calls a "Cursor for decision-making"—a system that learns from high-stakes real-world decisions (initially in finance and research) by capturing decision trajectories, user feedback, and post-hoc outcomes to continuously refine its recommendations via a "generalized post-training" loop.
Why it matters: Kando AI is deliberately positioning itself as a vertical application of a pattern that has reshaped AI coding—the high-iteration-rate, feedback-driven adoption loop that Cursor pioneered. In coding, the metric that mattered was code-acceptance rate rising from near-zero to 60-70%; in decision-making, Kando targets "decision-adoption rate + post-hoc reliability." This is a direct extension of the Cursor-like pattern into a new domain: financial analysis and research, where the feedback cycle is tight (days/weeks) and the value of correct decisions is high. The bet is that the same mechanics—chaining user decisions into a fine-tuning signal—can work in knowledge work outside software engineering.
The team brings a mix of frontier AI research (Wu was first Chinese recipient of Apple's PhD fellowship, former Tencent AI Lab lead on trustworthy AI) and finance domain depth (Mao, a Tsinghua/HKU grad, was a top-tier investment banker and founded a prior embodied AI company). The seed round is modest, but the ambition is structural: Kando wants to build a 'decision intelligence infrastructure' layer that sits between raw information and human action—a bet that the information→decision bridge, which has resisted digitalization for decades, is finally tractable with LLM-based systems that can absorb and learn from individual user cognition over time.
#KandoAI #AIWorkflow #VerticalAI #DecisionIntelligence #FinancialAI #SeedFunding #ChinaAI