Emma, a neural network, has been in development for the past year. For now this artificial intelligence engine learns from people who teach her accounting practices, how to value securities and other basics in finance. Shaunak Khire, co-founder at a company yet to be names, and partner at Magha Holdings, a company focused on technology and finance, sees other uses across businesses. He says she's not a bot, but can function like one.
Testing her knowledge, Stealth had Emma analyze Exchange Traded Funds (ETFs) and she came back with analysis and opinions on whether or not investors should buy them.
Emma works best when she can step in to analyze numbers for human-capital intensive tasks in retail, finance, and other sectors. Perhaps with the correct data she could analyze the best strategy for cross-device, cross-media advertising campaigns or targeting.
Here's how she works: Emma pools all available data from SEC filings, reporting statements, and management calls. She provides not only quantitative analysis but also cognitive analysis because she is being taught to interpret things in certain ways and extrapolate the analysis down to individual business units. She can analyze and give an opinion based on the data.
While Emma began in the world of finance, Khire sees other opportunities such as consulting and healthcare. She has a job as a consultant at a company in Japan where she gives advice to businesses, Khire said. And with nearly 35 billion patent visits annually worldwide in which 41% don't require a doctor's visit, per Khire, there's an opportunity to provide information about healthcare.
For some ailments, Emma can provide a diagnosis based on the person's physical conditions and suggest steps to lessen the symptoms.
People who are trying to self-diagnose themselves often search on ailments. Google health searches now include 900 medical conditions, and in Canada more than 17 million Canadians regularly browse online medical resources like WebMD, Healthline and Everyday Health each month, according to Google, citing data from comScore.
I see opportunities in business for manufacturing, advertising, and retail.
"We don't see a use for consumers right now, but we do see an application around messaging for businesses as the gateway to some of these services," Khire said. "The biggest challenge on the consumer side will become how bots rank and function.'
Khire said on average bots don't handle text well. Since they cannot analyze information well, many provide "weird responses."