About

AngelOne is a listed stock and mutual funds brokerage house in India. At the time of writing, AngelOne has a user base of about 18 million users growing over 30-50% annually and an average daily turnover of almost $400 billion.

In 2019, only 8% of Indian households held an active demat account, essential for trading in equities and mutual funds. The share more than doubled to 17% in 2023. The surge in retail investor base in India is attributed by experts to increased internet and smartphone penetration, attractive returns riding on the Indian growth story and digital first approach of most of the brokerage houses. 

Situation

AngelOne is experiencing rapid growth in its user base compared to the overall brokerage market in India. This has led to a significant increase in the number of incoming tickets for onboarding and support queries. With over 1,000 support agents, the contact centre now handles more than 500,000 tickets per month.

Previously, retail traders had to wait between 30 minutes to a few hours to receive assistance through email or phone. The resolution time was further impacted during events like IPOs, buybacks, and exchange glitches. Additionally, there was a disparity in knowledge among agents, with some being more informed about product features and troubleshooting than others.

Even the more experienced agents were not fully leveraging their collective knowledge, which ranged from publicly available resources like product documentation to user-specific sources such as user dashboard APIs and past ticket resolutions. As with any other product documentation, keeping all documentation up to date was a big challenge.

Solution

To achieve their goals, AngelOne and Alltius collaborated to reduce the contact ratio by at least 20% and decrease the median customer query resolution time from 2-4 hours to just a few minutes. They aimed to cover all possible information sources within the organisation, including static and dynamic sources, synthesised and unsynthesized sources, and structured and unstructured sources. They also wanted to reduce the new agent settling time from months to just a few days and decrease the information access time for agents from hours to seconds.

To start the process, a team of AngelOne support agents and product managers created a test Assistant with over 20,000 multilingual web pages that could process thousands of pages in less than an hour. They also created a test set of several hundred conversations that were monitored in Alltius testing Playground. After conducting a quality check of over 1000 data points and gathering feedback from over 50 agents, it was determined that Alltius was the best choice in terms of quality and capabilities. The assistants were then ready to be deployed across different user-facing channels via on-brand fully customizable web widgets.

During this time, an audit of Alltius' platform architecture was conducted by the infosec team to ensure robust security and business continuity measures were in place, especially for a partner of Alltius’ vintage.

Within three months, the assistants went live across web and mobile app platforms, increasing coverage from 5% to over 80% of the user base. The assistants were able to handle varied traffic patterns and scale up quickly when needed. They could handle peak capacity up to 10 times the median hourly conversation volume during critical events such as market opening hours, IPOs, and buybacks. AngelOne and Alltius held weekly meetings to assess performance and identify areas for improvement. The in-product audit trail and insights module helped the product team quickly understand documentation gaps and product issues. The custom labelling feature allowed for pattern recognition and insights in minutes instead of weeks. 

Through continuous testing after going live, AngelOne added several skills to their assistants, such as hierarchical search, answering date-time sensitive questions, detecting escalation intents, using option-based menu workflows, providing unique experiences for authenticated users, securely interacting with user dashboard APIs from the widget, and more.

Results

In just a few months, Alltius has already powered over 1 million conversations and recorded over 100,000 sessions per week. 

As a result, the contact centre has seen a 20% decrease in ticket volumes. Median wait time for resolving issues is now slashed to a few seconds from a few hours earlier. More than 65% of users are able to resolve their queries within 1 minute by interacting with the Alltius assistant, instead of waiting for hours for a resolution. This number is expected to increase even further as Alltius provides insights on recurring product issues and documentation gaps that are being fixed. 

This has allowed agents to free up more than 20% of their bandwidth to handle more complex and high-priority queries. Additionally, the product team now receives product-focused issue discovery, custom analytics, and insights within a day, instead of waiting weeks. This helps them quickly understand the distribution and recurrence of specific product issues, fix them and witness material progress subsequently in a matter of a few hours.

Way forward

As a result of the positive results in key areas, Angel One plans to expand the use of Alltius assistants in various roles. These assistants will be trained and coached with different skill sets to enhance onboarding workflows, assist support agents, facilitate conversational order placement, and support algorithmic trading. Furthermore, the user-facing assistants will be integrated with numerous additional APIs to provide instant support and improve self-serve assistance beyond the current 65% level.

Alltius is grateful to the AngelOne team for being a key design partner and early believer in shaping the platform.

‍