Could AI restore trust in banks?
Originally posted by Yolanda Redrup @ afr.com
US artificial intelligence company IPsoft wants to make a name for itself in Australia, going after the major banks and saying it can supply them with law-abiding "digital employees" that will help them rebuild trust following the royal commission.
To date the company, which has developed a "digital colleague" called Amelia designed to help employees in a wide range of industries do their jobs more efficiently, has flown under the radar locally, only employing five people despite appointing a local head two years ago.
But IPsoft chief commercial officer Jonathan Crane believes it is the right time to come out from behind the curtain and make a move in the region.
"There's huge pressure to find ways to build and enhance the customer experience and that's where we see Amelia and our cognitive capabilities taking hold," he said.
"All the banks here will tell you that after the royal commission report they're working on restoring trust. You do that through compliance and monitoring from a regulatory standpoint ... and in wealth management you need to set up accounts that have to adhere to new restrictions.
"The best way to do that is with a digital employee because they're constructed so they can't break away from the path. The expectations are set."
While the company would not disclose the names of its existing customers in the region, globally it works with the likes of Credit Suisse, Swedish bank SEB and US insurance company Allstate.
The business is one of the largest privately held AI technology companies in the world, with analysts estimating that it would generate somewhere in the range of $US700 million ($986 million) in revenue annually.
It has long been rumoured that the company is a likely NASDAQ initial public offering candidate, with these suggestions dating back to 2012 and re-emerging every few years.
While from the outside it sounds like the company's digital assistant Amelia would be designed to do away with jobs humans currently do, Mr Crane said it's actually designed to supplement humans.
In the financial services sector, Mr Crane said Amelia could be used to not only engage with customers on its own, but also to work alongside a human employee as a digital supervisor, keeping a record of all customer interactions. If a situation such as the royal commission occurred again, the data could easily be retrieved and audited.
"I think we'll see more and more use of technology in partnership with humans. We don't want to mistake AI implementation for job elimination," he said.
"Over the next five to 10 years we don't want to replace humans, but make them more capable, consistent and easier to train. It's a chance for us to step up and do more."
Unlike IBM's Watson, IPsoft's Amelia is designed to do menial jobs and work with humans, while Watson takes on complex data science challenges and can generate insights from data that humans would struggle to do in a timely manner.
Locally, IPsoft intends to keep its operations relatively lean but it is looking to form partnerships with the major consulting firms such as Deloitte, EY and Accenture, as well as more niche tech consultancies, to implement its products and drive adoption.
As well as in the financial services sector, it's doing work with government departments and in healthcare.
It hopes to be able to transform the way society approaches caring for the elderly, and also those with debilitating injuries that require rehabilitation.
If a person has to take 15 tablets a day, some of which need to be taken with food, and there's some that should not be taken together, Amelia can be trained to know what a patient should take when, what the possible drug interactions are and when to contact a person's doctor or nurse for help.
Emulating the human brain
In physical therapy it's already able to be used to monitor a person's progress doing at-home exercises via sensors, and it can send progress reports to the physician.
Mr Crane said Amelia's applications extended into so many fields because it was all built off one "basic brain".
"We're emulating the human brain, which is what differentiates Amelia from a chat bot," he said.
"We still want to enhance it with a faster learning ability, the ability to take analytics from previous interactions to guide decision-making and we also want to build up its emotional sentiment so that it can make more decisions that are linked to emotional behaviours, be it warnings from a medical standpoint, or consumer buying behaviour."
While quite different to a regular call centre chat bot, Amelia is an extension of this technology – the first of which was built by MIT in 1966.
Looking at Australia's position globally in AI, Mr Crane said it was behind nations like China and the US, but he was optimistic at the country's ability to catch up.
"The Australian government is looking at how AI can help improve the user experience of its services and government adoption is a very important aspect of overall adoption rates in any country," he said.