Woodside puts Watson on staff desktops

Woodside Energy is rolling out its implementation of IBM’s Watson engine to 50 staff for final evaluation ahead of an expected wider deployment later this year.

The gas giant – who, alongside ANZ Bank and the Immigration department, is one of the highest-profile users of Watson in Australia – revealed last year that it planned to use the software to power an internal program called ‘lesson learned’.

It appears the program has been rebranded to Watson@Woodside, although its purpose remains the same: to make it easier for staff to query 30 years’ of historical internal data.

Woodside has now revealed that this form of Watson will present to users as “a simple interface on [their] desktops”.

It is understood this will take the form of a desktop application, however it is unclear exactly how many of Woodside’s approximately 4000 staff will gain access to it.

A Woodside spokesperson declined to comment to iTnews about the reach or timelines of the project.

The company has spent eight months since June 2015 training Watson to meaningfully search Woodside’s data troves and present search results.

It is hoped that Watson@Woodside will “enable Woodside’s engineering teams to ask complex questions in natural language.”

“Allowing a broad population of employees to leverage this knowledge will enhance Woodside’s collective expertise in designing, fabricating and constructing major oil and gas facilities as well as managing major turnarounds,” it said last year.

While a small core data science team has led the charge, a “significant number” of employees from Woodside’s project function have “to load historical data from Woodside's major capital projects and shutdowns”, the firm said.

Others outside the project function are also understood to have spent time training Watson.

Woodside is also known to be scoping additional uses of Watson.

The next phase of deployment is likely to be used in the detection of geological and drilling hazards.

The company additionally hopes to tip more than engineering data into the Watson@Woodside implementation, enabling – for example – internal staffers to be able poll joint venture correspondence.