Kando AI, a portfolio company of Xinglian Capital, has completed tens of millions of RMB in seed round financing, aiming to build "the Cursor in the decision-making field".

Image source | AI-generated

Wu Bingzhe conducts at least one post-review of his decisions every day.

As a Doctor of Computer Science from Peking University, he is also a high-frequency investor. After the market closes every day, he reviews his judgments of the day: which contingency plans have been implemented, which have been disrupted by new information from the market, and which subconscious decisions have been proven correct after the fact.

This habit has lasted for many years. However, the problem is that most of the post-reviews have not been systematically documented and preserved.

"Your cognition is an asset, but right now all of it stays only in people's minds," said Wu Bingzhe. He and Mao Shuhan, CEO of Kando AI, are both active secondary market investors. The two share one thing in common: they both make high-frequency decisions in uncertain environments, and have realized that the cognitive patterns behind these decisions have never been digitized.

In June 2026, the two officially founded Kando AI, aiming to enable AI to continuously learn from real decisions and their outcomes, and build the next-generation self-evolving decision-making system.

Kando AI has completed tens of millions of yuan in seed round financing, led by Xinglian Capital, followed by Lihe Financial Holdings, with participation from well-known industrial investors. The Kando team currently has about 10 members, who come from leading domestic institutions such as Tencent, ByteDance, Meituan, and Huatai Securities. Wu Bingzhe, Founder and Chief Scientist, graduated from Peking University with a bachelor's degree in Mathematics and a doctorate in Computer Science. He was the first mainland Chinese recipient of the Apple PhD Fellowship, and former head of Trustworthy AI at Tencent AI Lab. Mao Shuhan, Co-founder and CEO, is a serial entrepreneur. He obtained his bachelor's degree from Tsinghua University and a master's degree in Finance from the University of Hong Kong. He worked at a leading domestic investment bank for many years, and once participated in the founding of an embodied intelligence company. He was selected into the 2025 Forbes 30 Under 30 list.

From Digitalization of Information to AI-powered Decision-Making

What Kando AI wants to build is not just a tool, but an environment where cognition can keep evolving. To sum up what Kando AI is doing in one sentence, Mao Shuhan, the CEO, said: "In the era of the Internet and mobile Internet, we have completed the digitalization of information; in the era of AI, what we want to promote is the self-evolution of decision-making.

"Over the past two decades, whether it's Tonghuashun and Wind in the financial sector, or paper databases, literature management tools and knowledge platforms in scientific research scenarios, what they essentially solve are problems at the "information level": helping you collect, organize, store, retrieve, and summarize information. However, the organized information, when left there, will not naturally turn into valuable judgments.

Whether in investment, scientific research, or other high-value decision-making scenarios, the truly critical step often takes place in the human brain. Researchers form judgments from massive materials, secondary market investors develop a "market sense" amid complex market noise, and scientific researchers form their own aesthetics, intuition, and trade-offs for a certain direction after long-term accumulation... These things are very difficult to express directly, let alone be simply digitized.

Mao Shuhan believes that there has always been a missing bridge here.

"Information and action are not automatically connected. The most critical link in between is decision-making, while the human brain, experience, subconsciousness, aesthetics and judgment habits have never been systematically preserved in the past. For the first time, AI makes it possible to build this bridge.

"This is also the biggest difference between Kando AI and most AI tools: many AI products still stay at the stage of "optimizing output quality" — making summaries more complete, retrieval faster, and writing more fluent; what Kando AI wants to optimize is something else: decision adoption rate and post-facto reliability, that is, whether users will actually adopt the suggestions given by AI, and whether the suggestions have truly entered the user's research, judgment and action process.

Mao Shuhan said that Kando is positioned as the Cursor in the decision-making field.

When AI coding tools first came out in 2022, the code adoption rate was almost zero. Today, Cursor's adoption rate has reached 60%-70%.

The reason why finance is selected as one of the verification scenarios is that this scenario is one of the most structured fields from the perspective of information dimension. Mao Shuhan said that in terms of dimensions such as feedback clarity, feedback cycle (at the day/week level), auditability, and decision value, the financial scenario has the highest overall score, making it one of the best fields to build a self-evolving decision-making engine.

Turn Products into a Continuous Post-Training Process

The product form of Kando AI is a workbench for high-value decision-making scenarios. But from the team's perspective, the workbench is only the product entry, and what it truly carries is a set of decision-making systems that can accumulate experience from real decisions, verify judgments, and keep evolving.

One of Kando AI's core capabilities is "memory". But Kando AI's memory is not a knowledge base in the ordinary sense, nor is it the same as common RAG retrieval.

Traditional RAG is more inclined to semantic similarity matching, while Kando AI's memory system is more like a recommendation and recall mechanism oriented to cognitive structure — what the system remembers is not just what the user has seen, but how the user understands problems, their preferences, and how they revised their judgments in the past, which is more like a kind of Harness.

"It will understand you better and better, but more importantly, it will learn from every interaction between you and the real world as your partner.

"Taking the scientific research scenario as an example, the product will form predictions, and the user will decide to adopt, modify, reject, or include them in the tracking list. After the results come out, the system will go back to the initial assumptions, analyze which judgments are verified, which variables have changed, and whether the errors come from missing information, reasoning deviation or implementation problems.

These decision trajectories will not stay in the chat records. The user's adoption, modification and rejection, the difference between expectations and results, as well as post-event reviews and attribution, will be organized into new learning signals, which will continuously update the long-term memory, task strategies, skill scheduling and model adaptation layer in the system.

This means that the results generated by the previous round of decisions will change how the system recalls information, organizes reasoning and generates suggestions in the next round, forming a recursive closed loop of "judgment — action — feedback — update".

From Kando AI's perspective, this represents an important change for AI products: in the past, the model's capability was mainly determined at the moment when the training was completed, while in the future, the intelligence boundary will be more and more determined by the experience accumulated after deployment.

Wu Bingzhe defines what they are doing as a kind of "generalized post-training".

In the traditional sense, post-training is more about adjusting model parameters; but from Kando AI's point of view, what really needs continuous optimization is far more than the parameters themselves, including how memory is stored, how to perform fine-grained recall, how to forget, how to design product interaction logic, how the context of skills evolves dynamically, and how different users' feedback can be organized into effective learning signals.

"You can regard the process of making products as a process of model post-training. The only difference is that in the past we optimized parameters, but now we optimize context, interaction and memory," said Wu Bingzhe.

This is also why Kando AI chooses self-research from the Infrastructure layer. Wu Bingzhe said that Kando AI hopes that all training links, feedback links and memory links are as transparent, monitorable and traceable as possible. What Kando AI precipitates is not a superficial layer of product functions, but a complete set of data flywheel that keeps growing around the real decision-making process.

In terms of commercialization, Kando AI chooses to start with users with high-value and high-frequency decision-making needs. The current seed users are mainly concentrated in professional user groups, including investors, scientific researchers, etc.

Kando AI is now in the internal beta stage. It provides services around users' continuous use, continuous memory and feedback, covering vertical industries such as finance and scientific research. From the team's perspective, any decision-making field with high value, non-standardization and strong cognitive dependence in the future may become the application target of this system.

"Kando AI is not just building a tool, but in those fields where there are no standard answers but are most worthy of being amplified, we will truly open up the closed loop of 'information → decision-making → feedback'," said Wu Bingzhe.

This is probably the most accurate positioning of Kando AI: what Kando AI hopes to finally build is a set of decision-making intelligent infrastructure that can continuously absorb human experience, understand real-world feedback, and continuously improve its own capabilities in every decision-making process.

Li Wenjue, Partner of Xinglian Capital, said: The next stage of AI applications is not just to process information more efficiently, but to enter the real workflow, learn how professional users make judgments, take actions and revise their cognition. Kando AI chooses high-value decision-making scenarios such as finance and scientific research, takes personalized memory and cognitive modeling as the base, converts users' research processes, adoption and rejection, and post-event reviews into learning signals, breaks through the boundary of one-time response of traditional AI tools, and forms a decision-making system that evolves together with users. We value that the team has both cutting-edge AI research, model post-training, system engineering and complex decision-making experience, and we also value its potential to precipitate proprietary workflow data from real use and promote continuous capability growth. If this closed loop is verified, what Kando builds will not only be an investment research or scientific research tool, but may become a new generation of cognitive infrastructure that connects information, judgment and action, enabling the most valuable judgment of individuals and organizations to be precipitated, calibrated and reused.