Organizer: momo, Chaincatcher
Guest: Vivien, MEXC Product Director
At Coinfest Asia 2026 held in Bali in August, the Alpha Arena S03 trading competition, sponsored by MEXC Ventures, had just concluded, where 20 traders from 12 countries fought fiercely in real-time, with a Japanese player making a comeback at the last moment to clinch the championship.
Outside the competition arena, MEXC Product Director Vivien Lin discussed another longer "competition" about how exchanges can catch the influx of users and funds in the wave of AI and RWA integration.
Before joining the crypto industry, Vivien Lin worked in foreign exchange and complex derivatives at Morgan Stanley. Her combined background in traditional finance and crypto allows her to understand the significance of this integration and transformation better.
She observed that with the introduction of perpetual contracts in traditional finance, the capital between the two markets has already been linked, and user behavior is also converging. When BTC prices are sluggish, capital within the crypto circle flows towards stocks and stock derivatives; when the market warms up, the capital flows back to the crypto market. Changes on the product side are similarly noticeable, where users inquire on MEXC AI about the differences between Nvidia and Intel, and hesitate over the authenticity behind tokenized stocks.
This change is also reshaping her product design logic. Vivien Lin initially wanted to create two different interfaces for two types of users but later realized they are more like different stages of the same group of people. Following this line of thought, MEXC refined its user profiles from a “one-size-fits-all” to a more layered “thousand faces” approach.
The explosive growth of AI is also rewriting users' trading habits and scenarios. In Vivien Lin's view, AI has transitioned from an "assistant" phase to a "co-pilot" phase for traders. Moreover, MEXC's AI layout goes beyond concepts, ranging from MEXC AI serving as an information steward to AI strategies capable of running back-tests, and AI search functions delivering all features with a single phrase, these functionalities have quietly woven into users' daily trading.
Recently, in an exclusive interview with ChainCatcher, Vivien Lin elaborated on how her traditional finance background shapes her product perspective, how to layer services for two types of users, the transformation logic between invisible product design and AI tools, and the changes in exchange product forms over the next three years.
I. From Investment Banking to the Crypto World: Changes in Product Perspective
1. ChainCatcher: You previously worked on foreign exchange and derivatives products at large investment banks like Morgan Stanley. How has this traditional finance background shaped your product perspective? Which traditional finance logics can seamlessly transfer to the crypto world, and which must be discarded?
Vivien Lin: Traditional finance provided me with excellent financial training, making me more risk-aware. I traded complex derivatives, needing to monitor Greek risks daily, many of which lacked readily available market hedging tools, which has made me extremely sensitive to even small data changes. In trading products, this sensitivity enables me to notice the tiniest points.
It has also kept me respectful of the market and regulation. Standing at the intersection of traditional finance and crypto, familiarity with regulations can bring different product insights; when designing compliant products, I better understand what regulators care about, and know how to meet regulatory requirements while preserving user experience, creating products that are both compliant and functional.
However, the pace of traditional finance does not work in the crypto space. Over there, developing a new product takes 6 months to a year, allowing for in-depth exploration of what users want and how the business model should work; in the crypto world, it operates 24/7, forcing you to rely on intuition about the market, user psychology, and the product itself to make quick judgments, and to rapidly rally resources from support departments and within the company, so everyone works together to quickly iterate on the product.
2. ChainCatcher: In the past one or two years, many traditional financial platforms have also been expanding into crypto, with “super accounts” becoming an integration trend. In contrast, what are the shortcomings and barriers of crypto-native exchanges?
Vivien Lin: Let's start with shortcomings. Early crypto market products were designed for speed, lacking segmented regulatory requirements, essentially offering one product to sell globally. Now, different regions have varying regulatory requirements, and users have become more diverse, including novices, high-net-worth individuals, professional institutions, arbitrage funds, as well as those coming from traditional finance who are not familiar with the crypto space but have strong learning capabilities. Whether they can solidify the infrastructure while accommodating user demands and the diversification of regional regulations will become a core test for exchanges to build their moat in the next 1 to 2 years.
As for barriers, crypto-native exchanges excel in retail products, as this market started with retail investors, and institutions have been absent for a long time. Traditional financial institutions understand finance and risk better, but when it comes to understanding retail investors, they fall short compared to native exchanges; the latter are built on a deep understanding of users, iterate quickly, and what traditional finance needs to learn is professionalism.
II. User Behaviors in Crypto and Traditional Finance are Converging
3. ChainCatcher: There are significant differences in trading interface and functional needs between crypto-native users and traditional financial users. How does MEXC balance the experience conflicts between these two types of users when creating a one-stop platform that includes stocks and ETFs?
Vivien Lin: The first principle of the product team is to provide users with a smooth experience. Based on this, the key is to observe what users truly need. I often cite two examples.
The first is copy trading, which is very much a crypto-native product. For veteran crypto users, it's an old play, whereas traditional finance users see it as a new tool; they need to find credible experts to trade alongside. So on one hand, we provide filtering standards to help users judge what types of traders are credible; on the other hand, we introduce more assets to give these traders greater room to operate. This is a fresh experience for traditional finance users.
The second is our newly launched Real Stock, which caters to the demands of traditional finance users. Crypto users care about how high the leverage is and how smooth the operations are, while traditional finance users primarily want to know if they are genuinely acquiring a stock when they buy one. Faced with the plethora of tokenized stocks in the market, they always have a question mark in their minds, which hinders their experience. After recognizing this point and discussing it repeatedly, we ultimately decided to directly connect with traditional brokers to gain real stock exposure. This technical path was chosen to address users' core concerns.
4. ChainCatcher: Real stocks and traditional financial perpetual contracts are key areas MEXC is currently promoting. Based on the current user data and product feedback, what are the behavioral differences between these two types of users on these products? What optimizations has the product team made based on these differences?
Vivien Lin: Research shows that the overlap between traditional finance and crypto users exceeds 60%, but our own experience has not reached that figure yet.
However, after the introduction of traditional financial perpetual contracts, the capital between the two markets has indeed been linked. Recently, when BTC was sluggish, a lot of crypto money flowed into stock tokens and perpetual stocks; these past few days, as the crypto market warmed up, capital has flowed back from stock positions to mainstream crypto assets.
With the capital linked, user behaviors are also converging. Two years ago, we began promoting the integration of traditional finance users and products with crypto. At that time, traditional finance users criticized crypto mainly for its opacity and conceptual nature. Many platforms launched index products or stock ETF types like QQQ, but they did not trust the authenticity of the underlying transactions and merely stood by to observe. Now, more and more traditional users proactively seek to understand how perpetual contracts are settled, and combined with our AI-assisted learning tools, the understanding cost is not high. Some friends of mine who are traders even said that this mechanism directly addresses the troublesome nature of leveraging in traditional finance.
We initially considered creating two separate interfaces for the two types of users. It was later discovered, however, that the users' mindset actually follows the attention to assets. When the AI sector performs well, many crypto users also ask on MEXC AI about the differences between Nvidia and Intel. Thus, the product logic shifted from serving two groups to serving different tiers of the same group, categorized by maturity such as beginners and experienced traders, rather than categorizing by asset class.
5. ChainCatcher: MEXC currently covers over 170 markets. How do you design layered experiences and growth systems for users at different levels?
Vivien Lin: We categorize users into two parts: those that can be perceived and those that cannot.
The non-perceptible part relies on more powerful analytical tools and big data technology. As soon as users enter the platform, we can determine which entry point they came from, if they are closer to the traditional finance profile or risk-averse, whether they are transactional or deposit-oriented. In the past, user profiles were very coarse, usually divided into three categories: novices, institutions, and VIPs. Many VIPs actually have operational habits very similar to institutions, but at that time, data and analytical tools were insufficient, causing the entire industry to have coarse granularity in understanding users. Now, profiles can be refined to a dozen categories; for example, a conservative user who only trades BTC and Ethereum and cannot tolerate significant drawdowns would be suggested AI strategies that focus on mainstream trends or reversal signals, with minimal drawdowns and back-testing support.
The perceivable part is the new VVIP system. Traditional VIPs are primarily categorized from 1 to 9 based on deposit amount and trading volume, while VVIP innovatively introduces the M Score. It does not only recognize users based on assets but also translates all behaviors after users enter the platform into scores, resembling a loyalty card. Daily logins, participation in activities, and trying new products all accumulate loyalty. The more loyal a user, the more new product experience opportunities, entrance and limits for high-interest financial products, as well as higher cashback ratios for MEXC cards will be offered. The entire VVIP system and user profiling system are linked to the M Score, with relatively transparent rules, which is currently one of the more advanced perceived layered products in the market.
6. ChainCatcher: Can you share a specific example of localized feedback? How did it drive the iteration of core functions?
Vivien Lin: In the past three months, we have invested a lot of effort into something that users may not be able to perceive, which is optimizing our internal operational tools. Previously, operations colleagues could only configure activities based on their personal understanding, which varied greatly; a ten-year veteran operator and a new operator, or someone born in the '80s and someone born in the '00s, have completely different perceptions of users. Covering users in over 170 countries and regions, with different ages and risk preferences, makes it nearly impossible to achieve precise operations through human effort alone.
We leveraged AI and big data to detail user profiles, then through data engineering, crystallized a set of atomized internal products that allow operational staff to quickly launch and configure activities.
Now, for different countries and partners, configuring an activity that aligns with current events and community characteristics takes less than an hour. Previously, crafting a proposal, producing images, and translating into multiple languages took two to three days for the whole process. For instance, when the market surged yesterday, transitioning from ideation to launch used to take at least two or three days, but now it can be completed within half an hour, allowing users to participate in the market at the right time, paired with promotional activities to enjoy the moment of growth.
7. ChainCatcher: Besides the visible aspects to users, what other invisible product designs have played a key role in MEXC's retention and conversion?
Vivien Lin: We have accomplished a lot below the surface. First, there is technological optimization; we are undergoing a significant system upgrade. After the upgrade, users will directly feel two things: the speed and smoothness of opening and switching the app have improved. Sensitive users might notice it's slightly faster, while most just find it much more comfortable, unable to pinpoint what’s better, but for the underlying infrastructure, it is a transformative effect. The new infrastructure can accommodate a larger trading volume throughput, ensuring the system does not crash in extreme market conditions and supports our future exploration of lending and options-type products. This provides psychological security for users from the product and research team.
Another type of invisible change involves user account systems and trading requirements that users might not perceive. For example, in AI strategies, what users see is that even in a deep bear market, the strategies you offer are still yielding decent returns; where do these strategies come from? Behind the scenes, a team of four or five has spent three months researching historical investment masters' trading strategies and styles, localizing them based on changes and volatility characteristics of crypto assets to refine the product.
III. In the AI Era, How Will Future Trading Scenes Evolve?
8. ChainCatcher: MEXC has launched several AI tools. Based on data and user feedback, what actual transformations have they brought about? Have they increased trading frequency?
Vivien Lin: AI products can be divided into several segments. Those that users perceive and have clear intentions, such as MEXC AI and AI strategies, are the main focus. MEXC AI leans more towards being an information steward, gathering global news and the technical indicators you need to see. Its design purpose is very clear, and we track how many users genuinely use it. Users vote with their feet; if they feel we've done well, they'll come to ask questions daily. If they consider it part of their daily information sources, we have more opportunities to recommend current popular cryptocurrencies and trading opportunities, making it more likely for them to see and ultimately convert.
The intent behind AI strategies is more direct, as users primarily build trading systems at the order placement stage. We mainly look at two indicators: one is the repurchase rate; if users enjoy it, they will use it repeatedly; the second is how many users are genuinely building strategies and running back-tests. This also deepens our understanding of platform users, how many are more professional, and how many simply take a glance, find it difficult, and leave. The latter needs to be differentiated as being a product issue or if they were never the intended target users.
There are also efficiency-oriented products, such as AI news, AI search, and the command-line interface we will soon launch, which are part of early preparations for the next generation of product experiences. Taking AI search as an example, it solves the end-to-end jump problem. Previously, if you accessed the homepage and wanted to find a financial product, you'd first need to locate the financial entry button, then scroll through a long list; looking for niche assets, you'd have to scroll quite a bit and might miss similar names; on mobile, with limited pages, after flipping through a few, users often decide not to scroll further, leaving many useful products hidden. Now, typing "BTC" in the search bar directly provides the three most relevant news pieces, current price charts, technical indicators, ongoing activities, and whether BTC has any financial or yield-generating products. A very small search box consolidates a wealth of material. This could not be accomplished through manual hard coding; now, the underlying components, interface calls, and content presentations are to some extent intelligently filtered by AI. When a user searches for a term, they can see almost everything related to it on this platform.
9. ChainCatcher: Based on the development of AI, can you share a representative future trading scene?
Vivien Lin: I understand the development of AI in three stages.
The first stage is the assistant stage, where you ask a question, and it answers one.
We are now entering the second stage, the co-pilot stage, which is a phase that follows closely. Under prior authorization agreements, AI understands what I am doing. It will observe my actions; for example, if it notices I'm prone to impulsiveness during a major market drop, the next time I feel impulsive, it will remind me: instead of buying a high-leverage product here, why not buy spot, or I have a high-interest deposit available; would you like to secure a definite yield first?
Looking further ahead, we will see intelligent operations that bring entirely new experiences. Currently, if we need to find a function, we have to open the app and scroll to the corresponding page; in the future, there will be a unified entry point, for example, embedding MEXC's helper into the mobile system, connecting to Siri. You could directly ask Siri: “Today BTC seems to be rising well, can you help me earn $1,000?” Siri would call our command-line interface, connecting to all back-end services, including back-testing, and assess based on my current asset reserves whether it's possible to earn that $1,000 today, then execute it directly, or if it's uncertain, provide me with three options to choose from.
This will be close to the highly intelligent world depicted in "The Matrix". The technical barriers primarily lie in computing power. What application layer companies can do is make the calls more efficient, cost-effective, and precise; while we cannot resolve hardware and infrastructure issues, once these infrastructure issues are addressed, the product experience will undergo a transformative change, which I believe we will see within a year or two.
10. ChainCatcher: Looking ahead to the next year, what are MEXC's core plans for AI exploration and feature upgrades?
Vivien Lin: From a product perspective, AI technology has permeated every aspect of our products and development. However, visible external products that users can perceive need to return to first principles: user-centricity. We do not need to show users that AI is utilized extensively on this platform, but rather aim to integrate our understanding of AI and users seamlessly into every small feature and piece of information. You may not realize you are using many AI functionalities; you just feel that this platform understands you better, and that’s the picture we want to achieve.
To reach that vision, many foundational tasks must be undertaken first. One is to make AI better understand assets, which requires continuous market learning; the second is to enable AI to better understand users, relying on user profiling and big data. AI addresses analytical issues, but tasks like data storage, organization, and standardization cannot be performed without significant human input to lay the foundation; only then can AI truly excel.
Product direction will also follow market trends. Recently, the market has been relatively sluggish, and users are less inclined to buy trend-based products; this is also why we are developing AI strategies. AI strategies include trend-based strategies as well as strategies to earn price spreads during bottom consolidations. The groundwork laid in technology and data has allowed product promotion to proceed relatively smoothly.
11. ChainCatcher: Looking ahead to the next three years, as Crypto integrates with traditional finance, what fundamental changes will occur in exchange product forms? What is the MEXC product team preparing for future trading scenes?
Vivien Lin: Our core focus is to effectively serve users transitioning from traditional finance to the crypto space. Customer service, business development, partnerships, and product strategy will increasingly focus on understanding what these users want in terms of assets and tools, as well as whether our technical infrastructure can be elevated to provide them with speed and a variety of experiences comparable to traditional finance. This will be the most important strategy from this year to next.
Under this strategy, the operations team will look into traditional financial assets, while the business development team will strive to acquire traditional financial users and channels. For the product team, a significant challenge is to fill the gaps for products that are very user-friendly in traditional finance but are still lacking in the crypto space. Major categories are generally available, but details are missing, such as iceberg orders, which are a popular order type in traditional markets, and some crypto exchanges have them while others do not. We need to quickly fill in these minor points, so when users from traditional finance genuinely arrive in the crypto space, we can meet their needs and serve them well.
免责声明：本文章仅代表作者个人观点，不代表本平台的立场和观点。本文章仅供信息分享，不构成对任何人的任何投资建议。用户与作者之间的任何争议，与本平台无关。如网页中刊载的文章或图片涉及侵权，请提供相关的权利证明和身份证明发送邮件到support@aicoin.com，本平台相关工作人员将会进行核查。
