Challenging IBM's Watson, DataRPM upgrades natural language interface for its 'Sherlock'

IBM’s Jeopardy-winning computer Watson has set a standard for machine intelligence. But now Watson has some competition: Sherlock.

Fairfax, Virginia-based DataRPM is calling up the image of the legendary detective in its announcement today of a major new release of its analytics system, which it says improves the ease-of-use for regular end users. As analytics tools grow more powerful, their use by business users who aren’t data scientists is emerging as a key battlefront.

[aditude-amp id="flyingcarpet" targeting='{"env":"staging","page_type":"article","post_id":1578372,"post_type":"exclusive","post_chan":"none","tags":null,"ai":false,"category":"none","all_categories":"big-data,business,","session":"A"}']

“Sherlock is smarter than Watson,” co-founder and Chief Product Officer Ruban Phukan pointed out to VentureBeat, adding that the fictional character may have been the greatest data analyst yet conceived.

DataRPM’s new version 7 “dramatically” upgrades the natural language interface that it launched last year, Phukan told us, enhancing the ability of an end user to type in a question in plain English by adding a new level of machine learning.

The system, Phukan said, automatically translates a natural-language question — such as “What is my revenue by location by gender?” — into a SQL query and determines the most appropriate data model to render the answer.

The answer — data-derived insights or predictive analytics — is delivered largely through visuals.

By contrast, he said, Watson Analytics — an offshoot of the Watson knowledge engine that took some human Jeopardy contestants to the cleaners — can only provide a similar experience by bundling the basic engine with several non-IBM software tools.

“For Watson Analytics to answer questions using natural language,” Phukan said, “there is a significant screening and customization” process to set it up.

“With DataRPM,” he said, “it’s like plug-and-play, with no upfront time required to get started.” The new level of machine learning, he said, makes the system smarter over time, because it is learning from user interaction and other inputs about what is wanted.

From DataRPM’s website:

[aditude-amp id="medium1" targeting='{"env":"staging","page_type":"article","post_id":1578372,"post_type":"exclusive","post_chan":"none","tags":null,"ai":false,"category":"none","all_categories":"big-data,business,","session":"A"}']

Phukan also noted that IBM Watson Analytics is only available in the cloud, whereas DataRPM can be used from the cloud or installed on-site. And, he said, there is “a huge price difference,” with DataRPM starting at about $100,000 yearly and Watson running about 10 times that. DataRPM reports that it currently has about two dozen customers, including Cisco.

Although the nearly three-year-old company is positioning Watson as its main competitor, Phukan also points to ThoughtSpot. “The difference [between DataRPM and ThoughtSpot,]” he told us, “is they are an appliance,” while DataRPM does not require a hardware purchase.

Other vendors pushing the boundaries on natural language interfaces include Birst and Upshot. There are “very few players who are focused on natural language interface” for analytics, Phukan said. “We’re the only one that offers the full stack.”