Woodside Petroleum searches for data value with IBM's Watson cognitive computing

The Australia and New Zealand Banking Group was the first major Australian company to announce a partnership with Watson, in 2013, when it said the system would be used to help its wealth management division with customer service. Last October the bank's global wealth chief executive Joyce Phillips said after some intensive "training", Watson was ready to be put to the test by customers.

The move at Woodside comes after the company impressed analysts last week with progress on ambitious cost and productivity improvement goals, which have already netted $560 million of $800 million in expected benefits.

Mr Gregory said Woodside had run an initial big-data project late last year, which it had used to design a program of predictive analytics for the organisation. In an initial trial run on relatively complex systems, it had to analyse the running of two of its LNG plants.

"We were able to detect an incident in the past that was very difficult to spot and then turn that into a prediction seven days out from anything happening, so we could essentially react and do something about it before a future problem occurred," he said.

"It really underpinned a lot of the vision about turning all of our data into predictive data, rather than just reactive data. We have a big bank of knowledge from over 30 years of operations; all that engineering expertise sits internally within the company, and is almost like a latent asset that we need to bring to the forefront."

Mr Gregory said that while it wasn't cheap to buy Watson, the investment was viewed as a long-term play with the expectation that artificial intelligence will become more and more important.

Initially the company is using Watson on an internal program called "Lesson Learned", which pulls together decades of engineering data from numerous sources including testing, projects and messages and makes it accessible to a wide group of workers. It is expected this will lead to faster decision-making, improved process flow and better operational performance.

Engineers will essentially be able to ask complex questions of Watson in natural language, and then follow up its answers with further questions as the computer remembers the context of the earlier question, like a human conversation.

In future it is hoped that the use could be expanded to help plan and analyse other areas, such as exploration of future global operations.

"Ultimately it could," Mr Gregory said. "I think the algorithms need to evolve a bit more. It is a few years away from being able to do things like that, which is why we view this as being a long-term partnership."

IBM Australia and New Zealand managing director Kerry Purcell said Watson was proving increasingly alluring to businesses across sectors. As well as ANZ, it is also used by Deakin University, IP Australia and Immigration and Border Protection, which uses it to analyse unstructured data sources such as news feeds and government reports to spot potential risks.

"In any area where knowledge professionals are working with an ever-increasing volume of data, Watson is fundamentally changing how these organisations operate, make decisions, drive growth and gain competitive advantage," Mr Purcell said.