The Challenge

One of Africa’s premium financial services groups that offers a broad spectrum of financial solutions both locally and internationally, needed a tool that targets the South African millennial market and makes financial advice freely available regardless of current customer status. In partnership with AWS, Swipe iX utilised various cloud services to build a tool that conforms to the above mentioned specifications, known as the robo-advisor.

As with all South African financial institutions, the brand had an obligation to treat customers fairly (TCF). The Financial Services Board describes TCF as follows:

"Treating Customers Fairly (TCF) is an outcomes based regulatory and supervisory approach designed to ensure that specific, clearly articulated fairness outcomes for financial services consumers are delivered by regulated financial firms."

Well versed on the benefits of using cloud and AI technology to improve business processes, customer service and ultimately extract more value both for the business and the customer, the financial services stalwart sought an agency who it could partner with to ideate, scope, build and implement a robo-advice proof-of-concept (POC). Traditionally robo-advisors are computer automated investment platforms, however in this instance, the company sought a rapid prototype that could prove the feasibility of offering customers access to fair and trusted financial advice in their own time, and at their own pace via text or voice using their mobile phone or computer.

Goals

Simplify and automate product recommendations using text and voice.
Build a customer centric solution that offers trusted advice and recommendations.
Showcase how to leverage PaaS services offered by AWS to rapidly prototype an idea.

Our hypothesis

Imagine a world where customers are able to engage with a financial advice tool in their own time, at their own pace using text or voice to ask product and life stage related questions. Now imagine them receiving real time responses tailored to their specific needs without once having to speak to a human consultant.

The Swipe team knew they had to create a platform-agnostic tool that provides a trusted and judgement free environment for new and existing customers seeking financial/life-stage advice. In other words, an always on tool that offers the right answer at the right time. A key consideration was the fact that the client has a long and reputable track record. South Africans have come to trust the brand and the advice its financial consultants offer. This reputation needed to be upheld and the mandate was clear: offer customers a tool that helps them make better financial decisions.

During the ideation phase, Swipe's origination point was to ask the following two questions:

How do we build a rapid prototype that uses natural language processing (NLP) to recognise intent, ask relevant questions, produce a personalised result and capture an active lead?
How do we capitalise on the power, speed, agility and cost effectiveness of cloud services?

The answer - AWS Lex.

As a service, AWS Lex it is extremely powerful. When configured, it can provide a more natural interaction (whether through voice or text) between a human and a computer/mobile phone. It is also “exportable” and the custom built solution can therefore be accessed via a browser (desktop or mobile) as well as through third party channels such as WhatsApp or Facebook Messenger.

The Solution

Through the use of our rapid prototyping workflow, the Swipe iX team built a customised NLP skill set, powered by AWS Lex, that can be accessed via a web browser (mobile or computer) as well as through Amazon Alexa. The time from brief to completion of the POC was 3 weeks with a further 2 weeks spent on expanding the POCs feature set. The result is an agnostic, customer-centric bot (text and voice enabled) focussed on providing accurate and non-biased financial advice for one of Africa's leading financial institutions.

The Robo-advsior was aptly called Lizwi, which means "to have a voice" in Xhosa. The POC focussed on 4 key questions to which Lizwi offers intuitive, honest advice that users can trust.

Tools

Machine Learning

Networking & Content Delivery

Compute

Database

Security Identity and Compliance

Storage

The Result

Tools such as AWS Lex afford teams the ability to build working prototypes in rapid time. In the case of Lizwi, the first prototype was ideated, built and presented to client in 3 weeks with an additional 2 weeks spent on extending its feature set for a Client Board presentation.

Despite this short space of time, the result is a robo advisor that can meaningfully answer questions in so far as they relate to funeral cover, retirement annuities, education, investment options and more, and produce tailored results based on a user's specific responses. Due to the nature of the architecture, of which scalability and cost was the key consideration, the tool can be extended to include additional features, to become “smarter” over time and it can be exported for use in third party messaging apps such as Facebook Messenger and Whatsapp as well as into voice enabled devices such as an Amazon Echo or Amazon Show.

View Next Project

M-Net Voting

API • Widget Toolkit • Ux • Ui • Big Data • Analytics

Robo-Advisor

App • Ux • Ui • Artificial Intelligence • AWS