As founder and CEO of Kyndi, Ryan Welsh maintains a high standard of company performance.
“Our goal is to make people one hundred times faster and one hundred times smarter,” he says.
The secret sauce for people of the future will be their work sidekick — artificial intelligence, in the form of a machine-driven sidekick. With it, workers will work faster and more productively by gleaning knowledge from hundreds of thousands of sources — text documents, web pages, e-mails, video — and have it delivered to them in a synthesized, meaningful way so they can make good decisions.
Core to this enhanced productivity will be natural language understanding like the kind Kyndi has developed. According to Ryan, “Kyndi borrows the positive aspects of good old-fashioned AI and borrows from the positive aspects of mathematical techniques to achieve a more flexible, robust artificial intelligence.”
Welsh says you can look at the “intelligent digital assistant” to see the limitations of a strictly mathematical approach to natural language processing. “When you ask Siri a direct question, she parses your sentence literally and really has no idea what you mean. There are subtleties to communication. When I started to look at applying mathematical, machine learning, connectionist approaches of AI to text, they weren’t achieving the level of understanding that was required to address enterprise business problems.”
“By combining advanced natural language processing, semantic technologies, and mathematical techniques borrowed from the quantum sciences, we’re better and faster,” he says. “Better in the sense that we do very powerful things with textual data that preserves context and meaning. And faster because the system can be trained faster, allowing us to get solutions implemented quickly.”
Ryan’s interest in artificial intelligence began from a practical, not a theoretical perspective. Prior to Kyndi, when he was working in the world of finance, he was often asked to read numerous financial contracts in a very short time.
“It simply wasn’t possible to review so much material in such a short period of time, which made me realize that if machine intelligence could review the material and get a true understanding of what was being said, it would turbocharge the performance of the human who needed that intelligence to make business decisions.”
“Now, just a few years later at Kyndi, we’ve had a customer tell us we reduced twelve months’ worth of work down to seven hours. So the dream I’ve had is already within reach. And we’re just getting started.”