XpoLog Ltd., founded in 2003 and based outside Tel Aviv, is pinning its hopes on its augmented search technology, which CEO Haim Koschitzky promotes as a major upgrade to how most large businesses troubleshoot crippling computer failures today.
The software starts with a relatively straightforward search function that can swiftly pore through terabytes of back-end computing logs to find indicators of problems. That's not unusual; Splunk and other companies have offered that tool for years.
But XpoLog's distinction comes from its augmented search, which analyzes the data in addition to searching, using context clues to weed out red herrings, visually illustrating patterns quickly — and then using the human's feedback to fine-tune its analytical capabilities.
The old Google-style searches simply aren't enough, Koschitzky told me, when the clock is ticking on real-time IT failures. As cloud servers, mobile devices and continually updated software have become commonplace, the sheer volume of log data that gets generated is too much for any human to make sense of, at least in a timely way.
"It's trying to find a drop of poison in Niagara falls, every couple of seconds," he says. "You need a real technology to analyze that data, because it's constantly updating."
It's another example of the emerging field of machine intelligence, best exemplified by the IBM Watson project, in which the software "learns" from human input to constantly upgrade its analytical capabilities.
Some of XpoLog's clients include 888 Holdings, an online gambling website operator and payment gateway provider BlueSnap. Koschitzky believes insurance and medical companies are ultimately their ideal clients because of the sensitive, critical nature of their data.
An annual license to analyze 8 gigabytes per day costs $6,400 — and many customers do much more — but in February, XpoLog released a free version for small projects that will analyze just one gigabyte per day. Koschitzky hopes small teams within large companies see the value of the tool and convince their companies to buy more lucrative licenses to use across the system.
The company, which is privately held by its founders with no outside investments, grew revenue by 60 percent last year, he said. Koschitzky would not put a number to his hiring plans, but XpoLog intends to build out full sales, marketing and engineering teams in New York as quickly as revenue allows.
Also in the near future, the company is planning an out-of-the-box version of its software for individuals. Bringing this advanced IT troubleshooting to lay people (or at least technically aware non-experts) is key to the sales pitch: Integrating both analysis and search into the same product would potentially allow a bigger group of people to solve IT problems.
"If you'd like to scale this to include people who are not data scientists, you need something much more automatic and intuitive," Koschitzky said. "And that's what we're trying to provide."
Ben Fischer covers local and regional business in greater New York City.