Let investment wisdom be invoked like code, Questflow bets on AI Finance Agent

Author: momo, ChainCatcher

In recent years, achieving financial equity has been one of the most important narratives in fintech, giving rise to a large number of influential companies and products. For example, Robinhood has compressed the account opening process to just a few minutes, Hyperliquid has made on-chain perpetual contracts no longer just a toy for institutional players, and in the era of AI Agents, even tasks like monitoring the market, placing orders, and adjusting positions can be handed over to Bots for automatic completion. The barriers have indeed been lowered repeatedly, leading many to feel that financial investment has finally achieved equity for everyone.

However, reality tells a different story. Ordinary investors remain the most vulnerable group in a complex investment environment. The problem is that the first three waves of innovation liberated your "hands," allowing you to open accounts and place orders faster, but the judgment skills hidden in the minds of top investors—such as "what is worth buying, why to buy, and how much loss should trigger a sell"—have never truly been democratized.

Recently, we noticed an interesting project attempting to change this. Questflow previously gained endorsements from top Web2 and Web3 platforms like Google, Circle, and Coinbase with its multi-AI agent scheduling platform. Recently, it launched a brand upgrade, positioning itself as an AI Finance Agent, aiming to become the Codex of finance.

In other words, what Questflow aims to do is transform the judgment skills of top investors into financial strategies that ordinary people can also access. It is not just another AI trading Bot that helps users lower operational barriers; rather, it seeks to enable ordinary investors to use verified investment frameworks like an investment assistant while maintaining control over their funds.

### 1. What is Questflow's differentiated path?

Currently, the application of AI in the financial sector is very limited and one-sided.

Specifically, players in the market are generally following two paths. One path is to create execution tools that automate processes like account opening, order placement, and position adjustment. While this approach lowers operational barriers, it does little to assist investors with the decision-making issues they care about more.

This has led to the emergence of another path, positioning itself as an AI investment advisory tool. By integrating large models, it generates research reports, signals, and suggestions for investors to assist in decision-making, but this path also yields unsatisfactory results in practice.

Where does the problem lie? According to Questflow, the pain points can be summarized in three layers.

First, good investment judgment is unstructured; it resides in the experiences, conversations, and notes of fund managers. It is difficult to generate good investment advice solely based on scattered single-point information using the integration capabilities of general large models.

Recent research by Bridgewater Associates and Thinking Machines Lab also indirectly confirms this point. Their custom models trained on expert data achieved an accuracy of 84.7% in financial information screening tasks, significantly higher than the 78.2% of general large models, and the inference cost was only one-thirteenth of the latter. This indicates that in the financial sector, expert judgment is more important than the model itself, but the problem is that this judgment has yet to be productized and distilled.

Second, data, models, and trading venues are separated from each other. After AI provides suggestions, users still have to manually operate on the exchange, resulting in a disjointed process.

Additionally, there is a lack of credible participation mechanisms in the market. Traditional copy trading only provides signals without logic; users do not know why a strategy buys, when it should sell, or how much drawdown they can tolerate, leading to blind following.

For professional strategy providers, the dilemma is equally apparent. They possess validated investment frameworks and risk disciplines but lack the infrastructure to transform these unstructured experiences into scalable products. Existing methods either involve issuing signals on social media or privately managing trades, which cannot sustain monetization or establish long-term trust.

By targeting the respective pain points of ordinary investors and professional strategy providers, Questflow has chosen a differentiated path to build a bilateral market. On one hand, it allows the judgment of top investors to be distributed at scale; on the other hand, it enables ordinary investors to access this information transparently and controllably.

In other words, Questflow's core positioning is not to use general AI to predict the market but to structure the market judgments of top investors—including investment logic, strategy frameworks, risk control disciplines, and exit mechanisms—through AI, transforming them into intelligible, subscribable, and cross-market executable agents. The source of Alpha remains human; AI plays the role of making this judgment replicable, monitorable, and executable.

Furthermore, Questflow aims to bridge the gap between "understanding the market" and "actual execution," allowing ordinary investors to gain top-tier judgment through AI while also having the hands of a trader.

### 2. How does Questflow build the "Codex of Finance" to achieve financial equity?

The bilateral market built by Questflow clearly targets specific pain points, and the model is straightforward. But the key is how to achieve this; there are three core questions: first, how to "extract" top strategies from the minds of experts? Second, how to match the supply side with the demand side? Lastly, how to ensure the safety of users' funds? Let's analyze them one by one.

1. How to "extract" judgment from the minds of experts?

As mentioned earlier, the knowledge in the minds of top investors is essentially a collection of unstructured experiences. When to enter the market, which indicators to watch, and how much loss should trigger a sell—this information is scattered in their notes, conversations, and intuitions, making it impossible to directly replicate for another person. The first thing Questflow wants to do is to make this implicit judgment explicit, thus creating a capability layer.

Instead of building a super model from scratch, it has constructed a financial driving system. At the model layer, Questflow connects to mainstream large models available in the market, allowing users to switch freely based on different tasks, enabling AI to call upon the most suitable underlying capabilities during analysis, research, or trading execution. The platform has also created a Benchmark to compare the performance of various financial models in real financial scenarios, giving users insight when selecting models.

However, research from Bridgewater Associates and Thinking Machines Lab also highlights another issue: while large models are important, if the model is a generalist, it cannot handle investments; investing requires specialists.

Thus, Questflow has created a skills layer, essentially equipping AI with professional standards in the financial domain. Users can directly install officially packaged investment methodologies or upload their own trading systems and research frameworks, allowing AI to analyze the market according to their logic rather than general logic. Coupled with plugins that connect to real-time information sources like news, market data, on-chain data, and social media, AI gains the sensory capability to continuously perceive market changes.

Finally, by connecting execution accounts to real exchanges, brokers, and wallets, AI transitions from theoretical discussions to actual read-write execution. In other words, a complete investment process in an investor's mind is broken down into four reconfigurable components: model cognition, methodological standards, data perception, and trading execution, which anyone can configure according to their needs.

2. How do good strategies find followers, and how do ordinary investors find good strategies?

Once strategies are structured, the next question is why professional strategy providers would be willing to provide them, and how demanders can find suitable strategies. Questflow has designed a distribution and incentive mechanism to motivate both supply and demand sides.

From the perspective of professional strategy providers, the core pain point is that judgment cannot be monetized at scale. Issuing signals on social media can only generate some traffic, while privately managing trades is neither sustainable nor capable of establishing long-term trust.

Questflow's approach is to provide them with a monetization channel. Verified strategies can be published as Funds (internal strategy portfolios on the platform, not regulated funds), charging fees based on management scale and performance. The more followers and better performance, the higher the income. This compels professional strategy providers to continuously optimize their strategies and transparently disclose performance, rather than just making a quick profit like in signal groups.

From the perspective of ordinary investors, the core pain point is the inability to find credible following methods. Signal groups and social copy trading in the market have opaque strategy logic and vague risk boundaries; followers do not even know what they are following. Questflow's approach allows them to filter Funds like browsing a mutual fund supermarket, selecting based on market, asset class, strategy type, and risk profile, understanding the investment logic before subscribing and following, and autonomously setting fund allocation and risk limits.

Both sides get what they need: the supply side has continuous monetization motivation, and the demand side has transparent and controllable choices, forming a complete incentive loop.

3. How can users trust AI to manage their money?

Even if the previous steps are successful, there remains a fundamental question: how can users trust AI to manage their money? For investors, this is a bottom-line issue; as long as control over funds is not in their hands, most people will hesitate.

Questflow's response is straightforward: the platform acts only as the brain, not as the vault. Questflow's control layer ensures that users always maintain control over funds, authorizations, and risk boundaries, with AI only operating within the user's clearly defined permissions, position limits, stop-loss conditions, and approval methods. Users can define configuration amounts, position sizes, risk limits, and execution permissions. Ultimately, the steering wheel remains in the user's hands, while AI drives within the designated lanes.

### Can Questflow become the "Codex of Finance"?

Understanding the pain points Questflow aims to address and its solutions helps clarify why it strives to become the "Codex of Finance."

The term Codex originally referred to legal codes and texts, but OpenAI has redefined it. OpenAI uses Codex to provide a "harness" for Coding Agents, supporting AI in smoothly completing tasks like code generation, debugging, and API calls. What Questflow aims to do is to apply the same logic to the financial sector; it has independently developed a dedicated "harness" for financial scenarios. With this framework, AI can not only understand investors' instructions but also analyze, decide, and act according to financial business norms, providing a new paradigm for the financial sector.

This vision is grand, and the future market ceiling is very high. Questflow currently chooses to enter categories like Crypto, perpetual contracts, and prediction markets, which operate around the clock. These markets have extremely high information density and volatility, making manual monitoring costs unrealistic, precisely the scenarios where Agents are most needed. If it can validate the closed loop of "productizing judgment" here, it can logically expand into other traditional financial markets.

However, whether it can succeed ultimately depends on the fundamentals. Questflow originally started as a multi-AI agent scheduling platform and has successively formed strategic partnerships with top Web2 and Web3 platforms like Circle, Coinbase, Google, Near, and CoinGecko.

Additionally, Questflow has received funding from well-known investment institutions such as CyberFund (an institution under Lido co-founder), Delphi Labs, Systemic Ventures (supported by Ethereum co-founder Joseph Lubin), Animoca Brands, HashKey Capital, and ElizaOS, indicating a certain level of accumulation in AI agents. The key moving forward is how it effectively reuses and iterates these core capabilities in the field of financial investment.

From the perspective of market dynamics, the opportunities are clear. Trading infrastructure is already mature enough, with no shortage of API ecosystems and liquidity, but most AI financial projects in the market remain at the tool level. Some attempt to predict the market with trading robots, some sell black-box signals, and others engage in pure copy trading. The closed-loop product of "judgment plus execution" is almost still a blank space.

If Questflow can successfully run this closed loop, it is expected to become the infrastructure for global investors to access professional financial judgment, allowing the market wisdom of top investors to be indexed, called upon, and combined like a code repository. Although it has chosen a more challenging path, if successful, the barriers will be even deeper.