Thomson Reuters will be using the $100 million expansion of its Toronto Technology Centre to simplify and transform knowledge work in areas like legal research, contract review and e-discovery, according to its vice-president of research and development.
In a keynote speech to a group of chief digital officers this week, Khalid Al-Kofahi said artificial intelligence (AI) and machine learning will be a focal points at the facility, which recently celebrated its first anniversary. Thomson Reuters said it was putting capital investment and a 12-year lease on the expanded centre, which will house up to 1,500 when it opens in 2021.
Al-Kofahi used the CDO Summit to show off Tracer, an AI tool developed to assist one of the best-known lines of business in the company: media. Tracer uses an algorithm to look for news on social media and categorize whether it’s an event (like a Lady Gaga concert), or news, like a M&A deal between two firms.
Next, the system analyzes whether the story is based on fact or an opinion, and if it’s the latter, whether it’s an opinion from someone who could be considered an expert. Tracer contains more than 720 features and has almost 80 per cent accuracy, Al-Kofahi said.
Tracer, however, is just one area where Thomson Reuters wants to make use of AI and machine learning. Al-Kofahi defined his team’s mission as “follow the business and lead the business,” which will require a combination of technical expertise and the right kind of information baked into the applications.
“At Google they have a saying that more data beats algorithms. I say better data beats both,” he said. “The question is, how can we transform from a series of transactions (between a human and a computer) into a more engaging dialogue?”
A Thomson Reuters client might use one of its e-discovery services, for instance, to find all the legal cases decided by a particular judge that focused on age discrimination over a 20-year period. That’s transactional, Al-Kofahi explained, because it’s largely a matter of classifying documents and bringing up a result. The more advanced forms of machine learning would be directed at open-ended kinds of questions, like, “Is Company X over-valued?” which would require highly sophisticated AI capabilities.
“That’s not a ‘find,’ that’s a research task,” he said. “We need to be able to transform that kind of task into an ask.”
Other examples could include tools that look for abnormalities in a contract, or ways to proactively detect events that could have an impact on supply chains, Al-Kofahi said. As the R&D team continues to increase its resournces, the company will be thinking more deeply about how financial advisors and other members of its customer base need to think and act.
“They engage in a journey of discovery to find information about a problem. They analyze it and discover some gaps. As their understanding matures, they make decisions,” he said. “That means our work is about transforming how professionals find, analyze and act upon information.”
The Toronto CDO Summit wrapped up on Wednesday.
Latest posts by Shane Schick (see all)
- Microsoft FutureNow gathers AI experts to map the best routes to enterprise adoption - March 23, 2018
- InsideView research reveals critical attitudes among Millennial sales and marketing professionals - March 22, 2018
- IBM Think positions ‘analytics governance’ as the key to enterprise data-driven decision-making - March 21, 2018