Building Robo Advisor

Problem: Signing up for an investment can be a lengthy, tedious process filled with paperwork. Potential investors often do not know which product is most suited to their needs, which necessitates scheduling an appointment with an advisor. The investment process also involves capturing client details; an extensive process of filling out forms which also commonly requires an inconvenient appointment with an advisor.

Solution: A digital investment advisor. A short conversational interface which captures key pieces of information to be used in the needs analysis algorithms. Advice, signup, and management can all be done online at any time, eliminating physical forms and advisor meetings.

Project overview

We worked with Nedgroup Investments alongside their international project partners, to deliver a digital “Robo” advisor. The advisor was to be integrated both with EEE for digital onboarding, and with the project partner’s sophisticated algorithms which provide investment recommendations.

The chat feature was to be embedded in their Extraordinary Life website and presented as the first step in investing with Nedbank:

Extraordinary life

Highlighted features

We wanted the digital advisor to include mechanisms to set it apart and make an often-tedious process feel exciting. To make capturing details more engaging, we incorporated elements such as an interactive facial analysis tool. This software ‘guesses’ your age rather than just asking how old you are.

For the first phase in developing an investment recommendation, we selected three pathways potential investors could choose from. These options were based on the most common reasons for wanting to invest: retirement, education, and a personal goal such as a holiday. Later, we added a fourth pathway to allow for a ‘just save’ option if the client had no specific goal in mind.

Other highlighted features which made the Robo advisor an interactive, efficient, and personalised solution to receiving investment advice include:

• Quick and easy capturing of salary, age, children’s names, and other personal details
• Assessing risk aversion based on questions the digital advisor asks
• High integration with Nedbank APIs and with our international partner’s APIs
• Personalised investment recommendations for specific and non-specific goals
• User location can be recognised and used to recommend schools
• An interactive map which can be used to pick schools
• School recommendations based on fee structure and needs
• Instant recommendations
• Ability to transition from chat experience directly into digital onboarding to sign up for the investment

Technology used

The following tools were used to create Robo advisor:

• ES6 JavaScript
• ASP.NET
• Google Location API integration
• Nedbank Nedsecure API integration
• Microsoft Face API integration for facial recognition
• Webpack
• SCSS
• CosmosDB
• Microsoft Bot Framework
• Application Insights
• Custom built webchat
• Microsoft Service Bus

We used ASP.NET, Azure and the Microsoft Bot Framework to build our chat bot because of the seamless integration between these respective parts.

Azure offers us easy ways to scale and maintain reliable services, with Bot Framework providing integration into various chat platforms and the ability for us to use our own custom JavaScript webchat.

Our custom-built summary dynamically updates as the user provides data about their financial goals and profile, which we use to keep the user up-to-date on their recommendations.

We have chosen to combine our own custom components with the built-in functionality provided by services such as Bot Framework and Azure in order to achieve a uniquely functional robotic advice experience for the user.

To keep the experience fun for the user and to provide the best advice possible, we integrated multiple external APIs (such as Microsoft’s Facial Recognition API and Google Location API).

Architecture