If Watson — the supercomputer first introduced via “Jeopardy” four years ago that’s become a key part of IBM’s business friendly story — was capable of all the great things IBM (NYSE: IBM) said it could do, then Hussey knew that data he had would find a good home. The trick would be to figure out which of Watson’s 30 cognitive APIs, each of which are trained to perform a different skill set and can learn to improve their performance, would best suit his company StatSocial’s needs.
So Hussey invited Kunal Kapoor, a business development manager at IBM Watson to come to his midtown offices in April and try to convert members of the StatSocial team that saw IBM as a “scary big box” in which their little startup might be lost.
“The truth is, there was a lot of resistance,” Hussey said during a conversation last week at IBM Watson’s headquarters on Astor Place in Manhattan’s East Village. “But what convinced me was when I did some testing.”
To determine which APIs are best suited for which startups, IBM engages in an on-boarding process that sounds a bit more like online dating than software sales. Since Watson’s APIs (application programming interfaces) give access to tools that learn and change over time, it’s doubly important to get the right match early.
Read original article

