AI needs a healthy serving of data to succeed

The rise of artificial intelligence and machine learning may be making today’s firehose gush of data more manageable, and open up more avenues for bringing out the value from these mountains of data. At the same time, AI needs all the data it can consume, requiring large data sets to function properly. AI and data need each other.

This point was recently brought out by Brigg Patten, writing in CTO Vision, who noted how AI has changed the game in terms of leveraging data insights. “As big data moves to the maturity phase, firms are now looking for ways to combine the scale of AI and the agility of big data processes to bring about an acceleration on the delivery of the much-needed business value,” he states.

Some insurance companies are leading the way in the AI data revolution. Patten cites the ongoing work of MetLife’s Pete Johnson, who states that data empowers AI in several ways. “Machine learning has scaled up,” Johnson is quoted as saying. “Data has made it possible for the development of scaled-up algorithms.” In addition, big data technology “has made it possible for organizations to process large volumes of data in relatively short durations.” Finally, with the availability of huge volumes of data, companies have access to “large datasets that were previously not available. Today, organizations and individuals can access transcription, ICR, image and voice files, logistics data and weather data.”

Those organizations “with the right type and quantity of data has the upper hand over rivals,” Patten states.

Vala Afshar, VP at Salesforce and an industry luminary, also connects the dots between AI and data in a recent article at International Business Times, describing the symbiotic relationship between AI and data. “The most common applications of AI in the business world are around using machine or deep learning to automatically analyze big data,” he states. “This means companies can now better understand what customers want or need, and then automatically deliver recommendations to improve their lives. Importantly, unlike the bots of our imagination, this is something that's available to all organizations today.”

To make the most of AI, Afshar urges organizations to move away from old, siloed data. “An important part of preparing to make the most of AI – and supporting the customer journey end-to-end – lies in data consolidation.”

Many organizations “are sitting on mountains of unanalyzed customer data,” says Afshar. “By the time they get round to entering it into their analytics systems and understanding it, it's often already outdated. However, AI relies on up-to-date internal and external sources – including cloud, social, mobile and the Internet of Things (IoT) – to arrive at meaningful and actionable recommendations.”

Businesses need to be ready for this new data, he continues. “This means moving away from separate, disparate legacy systems for different areas and embracing cloud-based technology where unstructured data from systems like IBM Watson can instantly be integrated into your own customer data.” Legacy systems also may complicate things, Patten says. “Traditional computer processors cannot process big data. Big data can best be processed by a GPU database, which has the flexibility needed to handle a significant amount of data of different types.”

AI and machine learning have been around for some decades now, Patten adds. “Lack of datasets of appropriate sizes prevented technologies that would provide meaningful learning and progress. The application of AI and machine learning in business has been facilitated by the ability to access big data.”