Consumers — even if they don’t know it — use artificial intelligence (AI) in many ways, through applications like Microsoft’s Cortana personal assistant or Google Photos, which uses machine learning to group together a user’s pictures. Although AI hasn’t yet taken off in the enterprise market, it soon may.
AI could be used to enhance customer service, provide companies with recommendations based on data analytics, root out fraud or help manufacturers find defects in products before they’re shipped.
The market for AI in the business world is going to heat up, according to research firm IDC, which predicts that the market for cognitively enabled applications and software is going to be worth $40 billion in 2020.
One kind of AI that has gotten a lot of attention lately is machine learning, an artificial intelligence that can learn from and make predictions based on data it accesses. Google is building the technology into the company’s cloud product in hopes of winning enterprise business.
Yet for all of the discussions about how AI is changing the world, it hasn’t transformed the business market quite yet. Stephen Pratt, a former Infosys executive who left his job after eight months of running the Watson AI project for IBM Services, wants to change the equation.
Consumers are getting a taste of AI via smartphone apps and predictive product engines from Amazon, but companies have only scratched the service. "But when it comes to optimizing a company's sales force or supply chain, that's just in its infancy," Pratt, Noodle’s CEO, told IDG News Service.
Pratt expects AI to become a normal part of how businesses get their work done. "I think AI will be the biggest competitive differentiator in the next three to five years," he says. "Executives who aren't using some form of AI to help them are very quickly going to seem outdated and out of touch."
AI is built on several foundational technologies, including natural language processing, speech recognition and structured data analytics, according to IDC analyst David Schubmehl, who specializes in AI. The artificial intelligence engine then uses those technologies and the data it is working with to create a statistical model of the data that can be used, thanks to techniques like machine learning, to make more intelligent or predictive recommendations.
AI is already being used in a variety of ways in the business world. Google uses AI in its AdWords advertising service and its core search, Schubmehl says. IBM Watson’s Engagement Advisor tool is designed specifically to enhance customer service. Saffron Technology, which Intel acquired in October 2015, claims that it used its AI to help a leading insurer save hundreds of thousands of dollars in fraudulent auto insurance claims. And Schubmehl says he just identified a use case in which IPsoft’s Amelia AI is being used to advise mortgage companies on what products to offer.
“I think there’s a tremendous amount of opportunity,” he says. “What we’re really starting to see is that organizations, as they invest in new types of software, are looking to make that software smarter.”
Incorporating various aspects of artificial intelligence can help businesses increase productivity, reduce manpower or provide advisory or recommendation services, Schubmehl says.
AI can help companies with financial fraud investigations, cybersecurity threat analysis, medical and health diagnostics or flaw detection, he adds.
While larger enterprises will likely develop products that use cognitive computing, Schubmehl thinks that small and medium-sized businesses will be “real beneficiaries” of AI. For example, he says that a smaller company could use AI to automatically schedule workers’ meetings based on the company’s scheduling program. He also notes that Microsoft has AI capabilities built into Delve, an Office application that allows colleagues to collaborate on projects in the cloud, surfaces relevant information and enables the sharing of expertise.
Schubmehl says companies like Google, HP Enterprise, IBM and Microsoft are already trying to make it easier for developers to connect to their AI platforms through application program interfaces to create services and apps. These AI platforms also require tremendous amounts of computing and processing power, which he says is good news for companies such as AMD, Intel and Nvidia that are investing in graphics processing units, processors that can assist in machine learning.