In the adviser market, ‘robo advice’ has been touted as a low-cost – albeit stripped back – alternative as spiralling costs, compounded by a decline in adviser numbers following RDR, has made the traditional advice process unaffordable for many.
The viability of these online investment management sites, based on algorithms that originally served the traditional advisory community, is being taken seriously by the City watchdog – which launched a joint consultation in the autumn with the Treasury to explore how they can improve access to financial advice.
Now a new form of artificial intelligence has entered the fray, promising to transform front line customer support for mortgage companies while helping them to allay dual pressures of driving costs down in a competitive environment while maintaining high satisfaction levels.
Its name is Amelia.
Named after Amelia Earhart, the US aviation pioneer and author, the algorithm-based system has been designed to be able to undertake the tasks of a typical mortgage firm’s call centre operative, including determining mortgage eligibility and providing product information, to allow them to take on more creative roles.
More than two decades in the making, the second iteration of the ‘virtual employee’ understands 21 languages, absorbs information from textbooks, transcriptions of conversations, email chains and just about any other written text in a matter of seconds.
Its creator, Chetan Dube, chief executive of tech firm IPSoft, said: “My wife thinks I’m having an affair because I have spent so much time with Amelia.”
He added: “People have been talking about the second machine age. The basic platforms on how services are rendered are shifting and Amelia is emblematic of that shift. Cognitive technology is going to be the trend of the 21st century.”
Call centres are typically the front line of customer support, but are often characterised by long waits and unwieldy menus, leading to unsatisfactory experiences and ultimately loss of business, Mr Dube said.
He added that traditional automated response systems, which are pre-recorded menus, equipped with speech recognition software that guides customers through a range of options, are cumbersome.
Amelia, however, has contextual filters which allows it to understand loosely stated problems and recognise sentences that have the same meaning but are structured differently.
When faced with foreign queries, the system will call upon a more experienced human agent to help resolve the issue. It will then listen in to the human-to-human interaction and create new steps in its process ontology, which will enable Amelia to address the same type of issue with subsequent callers.
Amelia is currently being used by an unnamed ‘leading’ UK bank to support staff who have questions on mortgage eligibility, according to Mr Dube, adding that the system has also been courted and tested by a number of multinationals in areas such as manning technology helpdesks and financial trading operations support.
Research conducted by analysts at US information technology research and advisory firm Gartner, found that five US media companies which implemented the system saw their mean time to resolve a telephone query reduced from 18 minutes and 20 seconds to four minutes and 30 seconds.
In addition, the average speed of answer reduced from 55 seconds to two seconds.
Gartner also predicts that by 2017, managed services offerings using cognitive platforms such as Amelia will drive a 60 per cent reduction in the cost of services.
Mr Dube said: “In this day and age, people are looking for instant response to queries. If this can be achieved through an AI system then I do not think that many people would object. The mortgage industry is looking more at automated digital solutions to further streamline the process of the preliminary validation of loans.”
As artificial intelligence evolves and becomes more sophisticated, the distinction between man and machine is becoming increasingly difficult to distinguish.
Just last year, a computer programme called Eugene Goostman became the first artificial intelligence to pass a test which investigates whether people can detect if they are talking to machines or humans by successfully posing as a 13-year-old boy.
The latest iteration of Amelia, which was unveiled in October, features new updates bringing Amelia closer to achieving near human cognitive capabilities, according to Mr Dube.
The system’s physical appearance and expressiveness have been transformed to create a more human-like avatar capable of deepening customer engagement.
Amelia, modelled on the human brain, can process information like a flesh-and-blood individual rather than simply mimicking human thought processes, according to Mr Dube, who said the system was equipped to sense human emotions and respond appropriately.
Mr Dube unveiled plans to go a step further by giving Amelia a physical ‘body’ by embedding its ‘brain’ into a robot.
He said: “Man needs to be able to communicate with somebody who understands, and is able to respond to his emotions – we have been exposed to dysfunctional automated AI for too long.”
He added: “Through the leverage of cognitive technology, we can create an efficient planet. We only use 10 per cent of our brain, which is a colossal waste. Technology can be a faithful friend to man and can help to create a more efficient ecosystem in business and answer environmental issues.”
The threat that machines pose to mankind is a concept explored by dystopian novelists and Hollywood film directors.
To root out ‘rogue learning’, every new piece of information that the system absorbs is subject to an overnight control process, enabling programmers to cherry pick what it should learn.
For Ray Tammam, IFA and mortgage broker at Manchester-based Fairstone Financial Management, the real threat posed by AI is to human employment.
He said: “Providers are faced with a moral dilemma – do they implement an AI system to save on costs, or should they keep people in jobs?
“I would probably rather speak to a human over the phone, but for straightforward queries, I would not be averse to speaking to an AI. Sometimes you will call up and the operator will give an answer which explores his past experience. I do not think that AIs would be able to do that – at least not in the near future.”
David Hollingworth, associate director of Bath-based London & Country Mortgages, said: “I do not think many advisers would object to speaking with an AI, as long as the information being given was accurate. Also, the responses mustn’t be generic. A human operator would either know the answer to a question or would have to go away to find the answer, or simply pass you on to a colleague. The AI must be robust enough to give equally sophisticated answers.”
He added: “The process can be very time-consuming. If the lender is busy, you can find yourself waiting on the phone for a substantial period of time. We will definitely see more attempts made by providers to innovate how they serve their clients.”
Myron Jobson is a features writer at Financial Adviser
Key points
A new form of artificial intelligence has entered the fray, promising to transform front-line customer support for mortgage companies
An IT company has devised a virtual mortgage assistant to handle queries
Questions arise over whether banks should implement an AI system to save on costs, or keep people in jobs
