Volatility has always been part of investing. What has changed is how quickly retail investors are now expected to make sense of it. A tariff announcement, an earnings miss or a central bank signal can trigger a rush of commentary, analysis and conflicting views within minutes. The challenge is no longer access to information but filtering it fast enough to act with confidence.
For instance, when US reciprocal tariffs took effect last August and equity markets lurched, investors on Tiger Brokers’ platform did not flood the market. Instead, they opened the app and started researching. Conversations on TigerAI, the AI assistant embedded in the Tiger Trade app, surged 1,378% y-oy that month. The number of monthly active users rose 618%. Crucially, the spike was not driven by passive browsing. Most queries centred on specific stocks and what the moves meant for positions investors already held.
“In moments like these, investors needed clarity fast on what was happening and what it meant for their holdings,” says Ian Leong, CEO of Tiger Brokers Singapore. That appetite had been building well before August. TigerAI launched in April 2023 and by the end of August 2025 had logged roughly 3 million total conversations and 300,000 global users, up 460% and 262% respectively from a year earlier.
Filling the gap
Since the launch of ChatGPT, investors have been drawn to general-purpose AI tools, and understandably so. Being able to ask a market question in plain language and get an instant answer is useful. The limitation is what those tools cannot see.
Such tools lack visibility into a user’s portfolio, live data integration, and connectivity to a brokerage’s infrastructure. As Leong puts it, general-purpose AI tools were not built for the moment an investment decision is actually being made, where timeliness is often the difference between a good call and a missed one.
To close that gap, TigerAI connects directly to Tiger Brokers’ internal financial databases and each user’s watchlist and positions instead of just using public information. This means the insights are specific to an investor’s portfolio. Each answer also includes clickable citations to original news sources, so investors can check the reasoning behind any response themselves.
That connection to real portfolio data is only as useful as the reasoning applied to it. Rather than a single model fielding every query, TigerAI uses discrete AI agents to handle specific tasks — such as options analysis, watchlist monitoring, stock screening — with each optimised for its function.
Powering that architecture is a combination of OpenAI’s GPT and DeepSeek-R1. The latter’s appeal lies in how it works through complex problems step by step, rather than simply retrieving the most probable answer. For investing, that matters when users are asking about options strategies, earnings implications or portfolio risk.
“DeepSeek-R1 operates alongside OpenAI’s GPT within our multi-agent architecture rather than as a standalone decision-maker. This layered approach ensures that the reasoning power of DeepSeek-R1 is matched by the guardrails a customer-facing financial product requires,” adds Leong. Within three months of integrating DeepSeek-R1, monthly conversations on TigerAI in Singapore rose 83% and active users grew 54%.
To ensure outputs are reliable, a verification agent reviews all responses before they reach users. This acts as a safeguard against the kind of AI errors that have made financial firms cautious about deploying these tools at scale. All processing takes place within Tiger Brokers’ own infrastructure, with trading data anonymised, encrypted and never shared externally.
What investors are actually doing with it
A better test of TigerAI’s value is how investors use it when markets are calm. While a sell-off can cause a surge in activity, steady use during quiet times shows if the tool is becoming part of investors’ regular research habits.
For example, in options — where the knowledge barrier between novice and experienced investors is steep — TigerAI has become as much a tool for building fluency as for surfacing data. By the time those investors speak to a human adviser, they tend to have a clearer sense of what they want to understand, which makes the discussion more useful for both sides.
How far an AI tool should go in guiding investment decisions is a question every brokerage building in this space has to answer. For TigerAI, every response carries a disclosure that its outputs are for reference only, and the reasoning behind each conclusion is made visible rather than buried. “Since TigerAI’s launch, we have been deliberate about this positioning. TigerAI is built to help investors cut through information noise, surface relevant context and understand the reasoning behind an analysis — but the investment decision should still remain with the user,” Leong says.
He is equally direct on accountability. “With over 300,000 global users and nearly 3 million conversations to date, every interaction is a touchpoint with someone’s financial life. Investors always make the final call but Tiger Brokers remains responsible for ensuring TigerAI’s outputs are transparent, well-grounded, and presented with appropriate context.”
Singapore’s investing reality, built in
One complexity TigerAI has had to account for is that Singapore investors rarely stay within a single market. Around 35% of Tiger’s active clients who trade US markets also trade on the Singapore Exchange (SGX), and many hold locally funded assets alongside international positions. For an investor tracking a US tech stock, a Singapore REIT and CPF-linked funds, separate tools can make the overall picture harder to see.
TigerAI covers SGX-listed stocks, REITs, and funds eligible under the CPF Investment Scheme and Supplementary Retirement Scheme, applying the same analytical depth to those portfolios as it does to US and Hong Kong holdings. “There’s no separate experience or reduced capability just because the capital source is different,” says Leong. This matters when market shocks cut across borders and asset classes. A single platform that can surface the impact across an investor’s positions gives users a fuller view of risk, rather than leaving them to piece it together market by market.
Knowing the investor
The next step for TigerAI is getting to know the investor behind the portfolio, which is considerably a more ambitious undertaking.
Investors differ considerably in experience, risk tolerance, and financial goals. Some pursue systematic fund investments targeting steady asset growth. Others are drawn to high-growth equities with a higher tolerance for short-term volatility. There are also those managing retirement savings where the cost of a poorly timed decision is not just financial but also temporal.
The same market event carries entirely different implications depending on where an investor sits in that journey. “The next step in TigerAI’s evolution is understanding each investor better — their habits, their preferences, their goals — and making the experience more personal. A sophisticated financial intelligence system has to read those differences and deliver insights aligned to each user’s actual needs,” says Leong.
Deeper earnings intelligence is also on the roadmap. Earnings season compresses a large amount of market-moving information into a short window. For most retail investors, the gap between the numbers and what they mean for a stock’s outlook remains wide. Bridging that, in plain language and connected to an investor’s actual positions, is where TigerAI’s next iteration is pointed.
As personalisation deepens, the boundary between research assistant and financial advisor will come under greater scrutiny from regulators and users alike. Leong is clear about where Tiger Brokers stands. “As personalisation deepens, more tailored insights mean users can make better decisions for themselves. The decision still sits with the investor, and TigerAI’s role is to give them clarity and confidence, not to replace their judgement.”
What last August showed is that investors will reach for clarity over instinct when given the right tool. Getting the technology to work was the first challenge. Knowing the investor well enough to make it count is the next one. “Everything comes back to one thing: protecting what matters to investors. With the power of AI, we’re empowering every investor, so confident investing is within reach for all,” says Leong.
