The AI Race is On Says Flamingo AI Founder
According to CEO & founder of Flamingo AI, Dr. Catriona Wallace, Flamingo AI is a “true” machine learning company that provides cognitive virtual assistants (called Maggie and Rosie) to large enterprises in the US, and Asia-Pacific. It eyes four verticals: insurtech, fintech, healthtech and telecom.
The Aussie company, which was established in 2013 and last year raised $15M, aims to solve the problem of low online sales conversion rates, and low levels of efficiency in call centers. Aside from being able to improve efficiency by ~30%, it also provides clients “unique datasets” that indicate what customers want to buy without any human involvement.
“We’ve studied what are all the traditional AI players doing that is failing and how do we come at it at a completely different way that is simple, it’s learning based and it is unsupervised. And this is getting great residence, particularly with these large enterprises that are having very very poor experience in dealing with some of these other traditional players” – Dr. Catriona Wallace .
Currently Chubb Australia, AMP, CUA, DirectMoney, MetLife Asia and Nationwide are some of its clients (that includes clients that are piloting the software as Flamingo only accepts paid trials). Specifically, Nationwide is using Rosie in its D2C channel for guaranteed retirement income to guide users through research, quotation, application and account setup for this product.
The company offers two main pricing models – (1) flat fee subscription model + pay per usage, or (2) revenue sharing, where they may take 20% of premiums closed via their virtual assistant.
Last, unlike other solutions that build their AI on top of third-party platforms such as IBM Watson and Microsoft, the Flamingo AI platform was built in-house and is completely independent. This gives the company full control – allowing for better scale and complete understanding of how and why the machine learning behaves the way it does. Another key difference between Flamingo AI and other solutions is the unsupervised learning model; “we can’t foresee the future, and nor can our clients. We ask our clients to predict what kinds of questions Rosie will be asked and about half the time consumer behavior is completely unexpected. Unsupervised learning let’s us adapt Rosie’s behavior in near real-time” – Flamingo AI’s Chief Data Scientist, Dr. Jack Elliot.