For a technology startup to really catch the eye of banks' top tech executives it must, first and foremost, meet a need that few others in the marketplace are meeting.
A case in point is LMRKTS LLC, a New York-based firm that aims to help banks reduce risk in their derivatives portfolios by using algorithms to spot excessive counterparty risk. The firm was showcased at a financial technology event in New York last week, and it left a lasting impression on many of the attendees.
"Though it has a relatively small customer base, LMRKTS is tackling a massive industry problem," says David Reilly, technology infrastructure executive at Bank of America. "I think their timing is good and the solution is compelling."
LMRKTS was one was one of six startups that beat out dozens of other financial technology firms for the right to present their stories to tech executives from 15 of the nation's largest banks at Credit Suisse's well-appointed building overlooking Madison Square Park.
The one-day event, sponsored by Accenture and the Partnership Fund for New York City, was the culmination of a 12-week program in which the firms were paired with senior bank executives and coached in how they could best meet the industry's needs. Other startups featured included Kasisto, which developed an app that responds to voice commands; Enigma, a firm that aggregates public data to help banks make credit decisions; and RevolutionCredit, which offers online financial education to adults with thin credit histories.
LMRKTS already has seven banks experimenting with its platform. It focuses on the derivatives contracts that represent the highest cost to banks, those signed outside the clearinghouses. Banks report and reconcile their counterparty exposures through these contracts on LMRKTS' platform, and LMRKTS' algorithms look at banks' exposures and suggest trades they may want to add or delete. LMRKTS gets paid if banks choose the exposure-reducing trades it recommends.
So why can't banks do all of this themselves? "They don't have a holistic view of exposures," says Lucio Biase, the company's founder.
Kasisto's app was also popular with the bankers. Its virtual assistant technology might best be compared to Apple's Siri in fact, the software came out of SRI International, which created Siri.
Like Siri, Kasisto's software is a combination of voice recognition, natural language understanding and artificial intelligence. It translates the customer's spoken words into text, uses natural language understanding to figure out what the customer wants, and searches through databases for the answers, which it translates back into speech.
For instance, a customer might ask, "How much did I spend at Whole Foods this month?" And the app will reply with a dollar amount.
"The idea of having a conversation with our mobile devices is not too far-fetched we talk to our cars and to interactive voice response systems," Nick Toro, director at Credit Suisse, points out. "The challenge is accuracy and context, having the system understand what we want to say."
Zor Gorelov, Kasisto's chief executive, observes that mobile is the fastest-growing channel for financial services, but that "there are fundamental usability issues that prevent people from doing more with their mobile banking app." Kasisto is testing the technology with two banks.
Standard Treasury also generated buzz on Demo Day.
The company takes historically proprietary systems and exposes them through an application programming interface layer that can be used by customers, by banks themselves or by third parties, to cut costs, streamline processes or create new revenue streams. A bank might use the software to fully automate the way it buys products from suppliers, for instance.
"APIs are a standard way two computers can talk to each other at high speed and with great autonomy," says Dan Kimerling, CEO of Standard Treasury. The company is already at $1 million in annual revenue and plans to open a New York office later this year.
Enigma, another startup in the group, makes public data readily accessible to banks that can be used to help make loan decisions. It has gathered public data from 100,000 public sources and provides APIs and data visualization to enable companies to make use of it. For instance, a bank might look up water usage records to predict the occupancy of a building based on its water use. This could be used to inform commercial real estate loan decisions. Another use case is employee benefit plans, which have to be filed with the Department of Labor. They can be used to figure out the number of employees, growth rate and turnover rate at a company, providing strong background for a business credit underwriting decision.
"We call ourselves the content arm of Big Data," says Craig Danton, head of business development at Enigma.
RevolutionCredit, another chosen startup, offers an alternative to credit scores for people with limited credit histories. Consumers who sign up for the service usually at the recommendation of their banks are put through a series of online courses and tests on financial topics, the idea being that they will be more creditworthy once they complete the course.
Traditional credit scores don't provide enough differentiation for the 120 million Americans who have middling scores, says Zaydoon Munir, founder and CEO of RevolutionCredit. And they do nothing for those with thin files.
"Consumers are getting stuck in a cycle of rejection," he said.
Through its tests, RevolutionCredit gathers behavioral data from consumers. At the end of the series, the consumer receives a certificate meant to prove they have the willingness and financial savviness to pay their bills. It's all free to the consumer, who is "invited" to participate; banks that offer this pay a licensing fee.
For one personal loan company, Munir said RevolutionCredit worked with 4,200 borrowers. At the end of six months, consumers who used Revolution Credit had 30% lower delinquency rates than their peers with the same credit scores, he says.
Another presenter was Pymetrics, a startup that uses gamification to help companies make hiring decisions.
Pymetrics co-founder Frida Polli says her company thinks of itself as a next-generation job marketplace that uses neuroscience and Big Data to help companies and candidates find each other. "LinkedIn meets Okcupid," she calls it.
Trying to find the right candidates is a challenge in financial services, Polli says. Typically 250 resumes are submitted for each position and about 60% of applicants are cut at the resume stage. About 30% to 50% of new hires don't work out after the first year, and 200% of salary costs are lost when a job turns over.
Pymetrics's games assess and profile each candidate.
Polli had the Demo Day audience play one of her company's games, one that looks at attention and impulsively. Attendees were shown colored circles and plus signs, and told to clap only when they saw a red circle. Several people also clapped whenever a blue circle appeared, showing their impulsivity. Some didn't clap at all, showing either lack of attention or lack of buy-in to the exercise.
The company works with 25 colleges across the country, including Carnegie Mellon, MIT, the University of Texas, Sanford, and UCLA, and has 25,000 users in its system.
Pymetrics has completed a pilot with one financial institution and has four others under way, she says.