“While the headlines in 2018 will be mostly about AI, most enterprises will need to first focus on IA—information augmentation,” said Ajay Khanna, VP of Reltio. The key is to get data “organized in a manner that ensures it can be reconciled, refined and related, to uncover relevant insights that support efficient business execution across all departments, while addressing the burden of regulatory compliance.”
Business context is also a critical factor in cognitive success, Khanna continued. “What many don’t realize is that the context of machine learning needs to be applied to your specific business and industry, with a focused set of benefits for each business user’s role in order for it to be accurately measured, so it doesn’t get labeled as yet another science project with limited value.”
Not Big Data
Ultimately, when done right and as efficiently as possible, the data behind cognitive computing need not be large volumes of “big data” either. “You hear a lot about big data, but small data is also driving real business applications for AI and machine learning,” said Ritika Gunnar, VP for IBM Watson and Data AI. “Small data is particular to an individual organization or a specific industry—it’s narrower and deeper than big data. For example, an oil and gas company’s engineers may have years of historical knowledge—everything from case files, incident reports on a particular rig, geological survey data—but it’s currently siloed with individuals or within separate systems. A key business advantage lies in tapping into organizational insights, historical customer data, internal reporting, past transactions, and client interactions. These elements are too-often underutilized. Plus, insights from small data can be coupled with insights from larger public datasets to give organizations the competitive intelligence needed to set themselves apart.”
There is no simple answer for determining how much data is enough to support AI and ML and, because the field of opportunity and potential is so broad, it comes down to what you are trying to do and what insights you are looking to get, said Watts. “For example, until Twitter recently closed off access to its Firehose—the full stream of tweets and other information—companies subscribing to this for trends and analytics would have been receiving around 600 million tweets per day which you could certainly classify as a lot of data. However, SAP and Salesforce customers range from small to enterprise and most of them will start to take advantage of AI for their CRM data, which for many will not be large quantities. It’s not so much about getting as much data as you can; it’s about working out what data might be valuable and pushing this into your learning datasets.”
Cognitive Apps
As they unfold, cognitive technologies will help open up a wealth of applications that were previously unavailable to enterprises. Mphasis NEXT Labs’ Ganesh reviewed some key areas where AI and ML are making a difference, including search and advertising algorithms; friend, movie, and book recommendation algorithms; driving-pattern recommendations; ?money-lending-related credit recommendations of peer-to-peer lending platforms; and the prediction of journey times to frequently visited locations.
Along with the enterprise data being gathered, this predictive power “gets enhanced as more partners—such as hotels, car rental companies, airlines, banks, and retailers—share user data. Enterprises are generating insights from multi-structured and multimedia datasets, the digital footprints of customer interactions, and customer intelligence across multiple channels, touch points, as well as social networks,” said Ganesh. Predictive analytics is also being leveraged for better planning, forecasting, and support for decisions concerning cross-sell, upsell, retention, loyalty management, risk mitigation, fraud detection, campaign management, and inventory management, he added.
Some learning organizations are seeing these technologies in action. Thomson Reuters, for example, recently launched a range of initiatives intended to support the reliability and trustworthiness of data and services being provided to its customers. The company’s Data Privacy Advisor, which is based on IBM Watson technology and supports data privacy professionals at multinational corporations and law firms, offers “a deeper understanding of the law and what their data privacy obligations are across multiple jurisdictions,” said Gunnar. The technology “allows the system to understand and interpret the context of the law and draw connections between relevant content related to a professional’s query.”
In addition, a service called Reuter News Tracer “helps weed out rumor and false news from social media feeds,” said John Hudzina, a researcher and architect in the Emerging Technology Group at Thomson Reuters. “Tracer mines potentially untrustworthy sources from Twitter for breaking events, the application applies cognitive computing to verify the event,” said Hudzina, noting that it examines the source’s social network and also attempts to determine potential biases. Hudzina added that his organization is relatively mature with its AI technology, as applied to its text-based information products—news, legal, and tax.