Deloitte Wins Hackathon Bid and Delivers AI App With OutSystems
Deloitte’s Risk Advisory team wanted to go beyond business-as-usual consulting engagements. So, when a large financial services institution hosted a hackathon for an unusual project, Deloitte partnered with OutSystems to win the contest. The institution needed a scalable solution to help clients monitor thousands of calls. Together with OutSystems, Deloitte built a working prototype of an artificial intelligence (AI) application in one week. And after winning the project, Deloitte delivered the finished front-end of the AI app in just five months.
App monitors 1000s of calls for risk
5 months: development time for front-end of machine-learning app
"Deloitte partnered with OutSystems because of the speed, flexibility, and adaptability their platform affords. Our clients can quickly change the BEAT app to meet specific needs, deploy use cases, and create workflows. All of this without having to rebuild the app or worry about the background processes."
Avoiding Risky Business
Challenge - Deloitte’s clients include financial institutions that must follow Financial Conduct Authority (FCA) rules, which include risk management. Many of Deloitte’s clients monitor customer service calls to prevent misconduct, complaints, conflict of interest, mis-selling or other related issues.
Given the sheer volume of calls an institution deals with each day, it simply can’t justify the time and resources it would need to spend manually monitoring every interaction. Deloitte quickly realized they could help their clients by making it easier to oversee countless calls without putting a massive toll on their employees.
"Originally, people didn't think we could do it. People thought this was not a problem that lent itself to technology."
An Unusual Competition for a Bid
Solution - A large financial institution approached Deloitte about competing in a hackathon with one of Deloitte’s rivals to develop a solution that enabled clients to monitor thousands of calls. Deloitte teamed up with OutSystems for the hackathon. Using the OutSystems low-code development platform, they developed the prototype for the given use case, integrating Deloitte’s Behavioural and Emotional Analytics Tool (BEAT) engine, an artificial intelligence (AI) platform in a few weeks. The client was delighted because they could immediately see what potential the application had for their use case.
Deloitte won the bid and used OutSystems to develop the front-end application in just 5 months.
The app uses AI and machine learning techniques. For example, acoustic and behavioral models are tailored to use cases. The result? The app can identify call segments that pose a conduct risk.
Andy Whitton, Partner, Risk Analytics says, “True Voice powered by BEAT listens to phone calls that clients have with their customers, and it doesn't just do the transcription of the words. It also listens to the sentiment in how the conversation happens. The app then uses AI to make a determination as to whether that customer seemed to be vulnerable and at risk of a bad outcome. So, the technology allows the client to home in on only the risky calls rather than randomly selecting any call. Our clients get a laser-focused group of calls for monitoring to help ensure that their customers have been treated fairly.”
The "BEAT" Goes On
Results – Several financial services organizations are now using BEAT to recognize conduct risk. Because BEAT is dynamic, intuitive, and easily personalized, it’s a popular app.
"Our clients can quickly change the BEAT app to meet specific needs, deploy use cases, and create workflows. All of this without having to rebuild the app or worry about the background processes. True Voice powered by BEAT replaces tedious work and drastically cuts costs."
Deloitte chose OutSystems because the low-code platform simplifies detailed processes so they’re easy to visualize, organize, and manage. In fact, one OutSystems developer incorporated existing technology into the app quickly and easily, freeing Deloitte to focus on improving the AI functions.
“With OutSystems, we evolved the True Voice app to where it is today,” says Asserati. “Now we have a blueprint for future apps that we can develop for other clients.”