IBM's Watson: From game shows to game changer in the enterprise
- 22 May, 2013 10:38
IBM's Watson artificial intelligence technology – perhaps most famous for its 2011 victory in US game show Jeopardy! – will be used by ANZ Bank to deliver faster financial recommendations to customers.
ANZ is one of the launch customers for Watson Engagement Advisor, a system that leverages IBM's Watson technology and can deliver an intelligent agent running on a single server that uses a combination of data about an individual client, natural language processing and a stored repository of information to offer tailored responses to customer queries.
Other customers include Celcom, IHS, Nielsen and the Royal Bank of Canada. The system can either be directly customer-facing or used to inform the client engagement of the legacy human workforce.
The new system can work across multiple channels, such as mobile devices or standard websites, and runs on a single IBM Power 750 server. This reduces the size of the hardware, IBM said, to the equivalent of a stack of four pizza boxes.
The original hardware used by Watson for Jeopardy! was a cluster of 90 Power 750 servers with 2880 processor cores and 16TB of RAM. The system used Hadoop for massively parallel processing and the GPFS for file storage.
The new Watson Engagement Advisor can be used for proactive engagement with customers, IBM said, and can also learn from past interactions.
According to IBM, the Watson system's performance has been boosted by 240 per cent and its physical requirements reduced by 75 per cent.
"We envision Watson to serve as cognitive assistant, adept and quick at making sense of big data, that can empower our regional bank advisors to better serve our 2 million wealth management clients across the region," ANZ's Joyce Phillips, CEO Global Wealth and Group Managing Director, Marketing, Innovation and Digital, said in a statement.
"We are pleased to explore with IBM how Watson can enable smarter, faster financial recommendations – yielding a customer experience that is simple, personalised and steeped in data-informed insights."