Turning opportunity into reality
nViso had already developed the innovative facial analytics algorithms it needed to build a solution for the financial services sector; the challenge was to find a way to deploy those algorithms on a massive scale.
To support the world’s major financial institutions, nViso’s new EmotionAdvisor solution would need to be capable of capturing and processing unstructured video data from hundreds of thousands of user sessions, integrating a wide variety of contextual data (which might be different for each institution or each group of users), analyzing the results, and returning personalized reports to both the user and their financial advisor.
Tim Llewellynn says: “Interpreting emotion is very contextual – there are so many variables that you can’t rely on a single, rigid data structure to handle all eventualities. That meant that a NoSQL database would work best for us, and IBM® Cloudant® was the ideal candidate.
“Cloudant’s document store approach means we can capture practically any type of data and store in JSON format. This gives us the flexibility we need to get prototypes up and running very quickly, and iterate to turn them into production-ready solutions for our clients.
“In a traditional project you might spend six to eight weeks just discussing the data structure to use, before you can actually start working on the application logic and the user interface – and when you’re dealing with line-of-business executives who need to show results in less than three months, that’s a big problem.
“Cloudant allowed us to shortcut the data modeling part of the development process, so it only took us eight weeks to build the entire application. Having fewer worries about the data means we have more time to spend on more iterations, and results in a better product at the end of the day.”
Seamless scalability and hybrid cloud options
IBM provides Cloudant as a database-as-a-service, running on SoftLayer®. This means than nViso does not need to manage the underlying infrastructure, and dealing with sudden spikes in user demand is not a problem.
“This is critical in consumer-facing applications, because traffic can be so unpredictable,” says Tim Llewellynn. “For example, we recently did an unrelated project for a media company where we had 1 million users over ten weeks – but 300,000 of them came in a 24-hour period when a company spokesperson mentioned the app on Reddit! Cloudant and SoftLayer take away those worries about scalability. They let us focus on building the best possible application, because we’ve already got the best possible infrastructure.”
While the cloud deployment model is critical for most of nViso’s current use cases, the company also sees Cloudant Local – a packaged version of the database that clients can run on-premise – as a key feature.
Jason Heath comments: “Financial services companies are strictly regulated, and there are often restrictions on how and where they are allowed to store certain types of data – particularly customer information. Cloudant Local gives us the ability to keep the most sensitive data housed on-premise if necessary, while still connecting to other Cloudant data stores via a hybrid cloud model. So our clients can easily prove to their regulators that our solution provides all the relevant safeguards.”
Transforming the way people feel about their finances
With the EmotionAdvisor solution now available online, nViso is able to offer the financial services sector a set of capabilities that it has never had before. Financial institutions can offer their clients instant access to a self-discovery session, which they can complete from the comfort of their own home, or even via a mobile device while they are on the move.
The client simply turns on their webcam or device camera, answers a few simple questions about their life situation – for example, their age and marital status, whether they have children, if they own a home, what their retirement plans are, and so on – and then watches a brief video that is tailored to be relevant to the answers that they have given. The video asks them to think about various financial scenarios, and analyzes their facial expressions as they do so. When the video ends, EmotionAdvisor directs them to a report that summarizes its findings, clarifying their emotional response to each scenario. At the same time, the solution also prepares a report for the client’s financial advisor, highlighting useful areas of discussion for their next meeting.
“When we ran a prototype of the solution for a large bank in New Zealand, over 200,000 people participated,” says Kyle Sleeper. “That’s more than six percent of the country’s adult population. It showed us that there is a real appetite for self-knowledge around personal finances, and people respond positively to the opportunity to learn something new about themselves and their emotions.”
After the project was finished, the bank found that there was a significant improvement in the way it was perceived by customers: it rose from fifth place to second in the national rankings in terms of being considered a good place for clients to manage their money.
For nViso, the success of this initial project showed that the financial services market had an appetite for emotion analytics, and the company is now seizing the opportunity to provide a new type of end-to-end service to clients throughout the sector.
“We’re not just saying ‘here’s a cool facial analytics API for your IT team to work with’,” says Jason Heath. “We’re saying that we can work with you to design a solution tailored to your needs, and deliver it from end to end within a few weeks. That’s a much more compelling proposition, because it means we can provide value to the line of business with much greater speed and agility than they can get from their own internal IT team.”
Tim Llewellynn concludes: “Our partnership with IBM empowers us to open up our ground-breaking emotional analytics to a much wider customer-base than ever before. EmotionAdvisor is just the first of many applications in our pipeline; Cloudant will help us rapidly develop and scale these solutions to benefit companies in many other industries too.”