City, watch out: the new codebreaker is coming

A Cambridge company is emulating Alan Turing with tools for detecting and predicting corporate crime and problem gambling

Hot on the heels of box-office hit The Imitation Game, about the Cambridge maths genius Alan Turing building a machine to crack the Engima code, comes news of another university first. This time the boffins at a young Cambridge company, Featurespace, are leading the world in “adaptive behavioural analytics” – a form of artificial intelligence that predicts human intention by analysing online data and so makes it possible to track down swindlers in any online scenario from gambling to rogue trading.

What’s more, says Martina King, chief executive of Featurespace, her team have been working with the Responsible Gambling Trust (RGT) to see if it is also possible to identify “problem gamblers” such as those hooked on the controversial fixed-odds betting terminals known as the “crack cocaine” of gaming. “Our research will significantly enhance our understanding of these machines and the way people use them,” she explains. “It’s an excellent example of how behavioural analytics might be used to tackle social, and commercial, challenges.”

You bet. Problem gambling is political dynamite: around 73 per cent of the adult population gamble each year, and around 450,000 are regarded as “problem gamblers”. A growing number are addicted to the fixed-odds terminal, which are increasingly popular in high-street betting shops.

The machines not only keep gamblers coming back for more but keep the money flowing out of their pockets: punters can stake a maximum of £100 every 20 seconds, allowing them to play up to £300 a minute. The terminals also take debit cards, and many gamblers have lost thousands of pounds in minutes – while some have lost their homes. About £1.6bn will be eaten up by the machines this year.

The Government is looking at controls and Ed Miliband has been more emphatic, saying councils will be given the power to ban the terminals if Labour comes to power.

Indeed it was the RGT, the UK charity dedicated to minimising “gambling-related” problems, which was so worried by the spread of these machines that it commissioned Featurespace to come up with ways of distinguishing between harmful and non-harmful activity using data from gaming machines. Its findings, said to be ground-breaking, will be published on Monday and then passed on to the Government for action.

Ms King was hired to run Featurespace two years ago from Mike Lynch’s Autonomy spin-out – the augmented reality start-up Aurasma – to help turn its complex algorithms into commercial reality. You can see why: attractive and warm, she is one of life’s born saleswomen. After selling and then running display advertising for Guardian Media Group, she went on to head Capital Radio, doubling sales, and then ran Europe for Yahoo! She brings gravitas, too, as a non-executive director of Debenhams and Cineworld.

We meet at her HQ in one of Cambridge’s bleaker technology parks. It’s a utilitarian office where more than half of the 29 staff have PhDs and are experts in “machine learning” subjects as diverse as Xbox games and writing code for high-precision equipment used by surgeons in operations.

Ms King says: “ I fell in love with these guys the moment I met them and wanted to help them grow the company from gawky teenage-hood to adulthood.” Her love has already brought maturity to Featurespace: since she joined, revenue has tripled, there are 21 paying clients and it is close to breaking even. Her latest coup was to persuade the former chairman of Capita, Gordon Hurst, to chair Featurespace.

Like so many Cambridge game-changers, Featurespace was born in an ivory tower – in this case the university’s engineering department – and is the work of the late Professor Bill Fitzgerald, an applied statistician, and Dr David Excell, now chief technology officer. (They also worked for the secret services on whether it’s possible to predict terrorist activity from watching passengers at airports).

They came up with their behavioural analytics engine, known as ARIC, after Betfair approached them some years ago to see if they could help improve the online betting platform’s fraud systems. Betfair liked the software so much that it led, laughs Ms King, to a Victor Kiam moment. “ They are clients and a shareholder.”

Now ARIC technology has been adapted for the financial services, insurance and retail industries and is being used by clients such as IG Index and Callcredit to detect and predict fraud in its many guises, from money laundering to the theft of credit cards. The latest collaboration, with accountants KPMG, is a humdinger: it offers clients in investment banking the tools to predict and model behaviour among traders and research analysts; the City had better watch out.

What next? Ms King says Featurespace has the potential to be as big as Autonomy or ARM – two of Cambridge’s brightest hi-tech stars. “The potential to apply our technology to organisations in every sector around the world is limitless. We receive calls all the time from clients that see how our technology could be applied to their particular business.”

So far there are no plans to list and raising new funds has not been a problem; Betfair, Imperial Innovations and the innovation charity Nesta are loyal investors, as is Mr Lynch, who is a non-executive director. Ms King will only say that ‘the man’s a genius” when I ask about Hewlett-Packard’s controversial $11bn takeover of Autonomy, over which Mr Lynch is being sued for allegedly inflating the accounts.

Being part of the Cambridge scene is a huge bonus, she says. “The network gives us access to today’s inventions that will change the world tomorrow.”

It’s also a network that needs careful nurturing. Ms King hosted the premiere of The Imitation Game at the Arts Picturehouse recently, to which some of the town’s finest minds were invited. She asked the Picturehouse to take extra care during the screening. “If anything went wrong in the cinema, the future of the UK would be put in doubt as some of the world’s greatest computer science brains were there.”

Martina King: Curriculum vitae

Date of birth: 7 March 1961

Education: Bonus Pastor school, London

Career:

2012 – CEO, Featurespace

2011-2012 CEO, Aurasma

2004-2012 Chairman, Radio Advertising Bureau; non-executive director at Capita Group, Johnston Press and Debenhams

1999-2004 European managing director, Yahoo!

1993-1999 Managing director, Capital Radio

1983-1993 Head of display advertising, The Guardian.

Personal life Married with two children

Favourite book A Christmas Carol

Favourite film It’s a Wonderful Life

Favourite music Anything I can dance to

Favourite holiday Activity based – skiing, walking etc

Favourite car 1983 Land Rover

Typical day “Wake early, work hard, hug the family on the nights I’m home, work some more, bed.”