AI seeks to analyze trader behavior to push profits, but experts urge caution

The use of artificial intelligence (AI) driven tools in stock-broking isn’t particularly new anymore, but a new branch of the so-called 'robo-advisory services' segment is emerging. These new firms, like Bengaluru-based Anastrat, don’t provide market advisory. Instead, they use the power of AI to analyze historical data, trading patterns and more to help guide traders into making their decisions.

Traditional robo-advisory services, examples of which include the likes of Sharekhan Neo, Angel Broking ARQ, 5Paise Auto Investor and more, use algorithms to offer automated investment plans. They typically use preset market strategies, combined with analysis of market trends to offer users ways to invest their money. On the other hand, firms like Anastrat offer suggestions like holding on to a stock for longer, diversifying portfolios etc.

Mohit Golecha, founder and chief technical officer (CTO) of AnaStrat, claimed that the firm is one-of-a-kind in India right now. While that might be so, and such firms are at a nascent stage, there are global examples like Polish data science consulting firm Sigmoidal’s deep learning-based investment strategy tool, Chicago, US-based Neurensic’s trade pattern analysis platform, and Seattle-based Kavout’s ‘K-Score’ – a service that uses data analytics and a predictive AI model to rank stocks by recognizing the investment pattern of various users.

Golecha said that the firm started offering public access in April and already has 110,000 registrations and 35,000 daily active users who use its platform. It has so far partnered with four brokerages in India – Zerodha, Fyers, IIFL and Dhan. 

To use the platform, users need to login through these brokerages, which allows the firm’s proprietary algorithms to access a trader’s historical data and logs of all trades they have done over time. The algorithms analyze this data to give users a score out of 10, with a lower score indicating that a trader has made more wrong calls on the market.

The platform's machine learning algorithms go through the trader's past data, and produce the score based on how they have held stocks, how quickly they have sold, and so on. It offers a 'traits' section that projects what a trader's market risk appetite is, and a separate 'insights' section, which suggests how trading behavior can be altered to increase profits.

According to Bhumik Gada, a trader and financial advisor, who works as a sub-broker for multiple brokerage firms, if platforms like this can provide reliable analysis, it could be a valuable tool for traders like him. Gada handles investments made by over 100 clients across India, and said it’s often difficult to follow the trading patterns of each client, which is what drives them towards analytics tools in the first place.

However, the efficacy of behavior analytics platform like this are yet to be tested fully. Mathew Chacko, founding partner and head of technology, media and telecom practice at law firm Spice Route Legal, said that AI tools that analyze trading behavior to offer market suggestions fall under the robo advisory segment, where regulations are not clearly defined in India.

“Firstly, there are no laws in India that tell customers what would happen if these platforms offer the wrong investment advice. More importantly, these advisory firms are not regulated by the Securities and Exchange Board of India (Sebi), and therefore do not provide binding investment advice to investors,” he said.

Further, Chacko said that analyzing historical data to predict future outcomes in a “volatile space” like the equity market is a risky factor – something that may not be particularly valuable for individual traders. “There are multiple factors that influence markets everyday, each of which fluctuate. Using such AI tools requires traders themselves to be veterans, and have a sound understanding of market risks already. Otherwise, there could be losses,” he added.

Moin Ladha, partner, corporate and commercial practices at law firm Khaitan & Co, added that the lack of regulatory definitions for AI and automated tools in India mean that such platforms could also come under the ambit of Sebi’s regulations eventually. “The way investment advisories and research analysts are defined by Sebi are too wide, and if you are offering any form of financial advisory services, you might easily fall under this ambit,” he said.