With AI widely predicted to be a major disruptive technology over the next five years and beyond, business leaders are taking a hard look at how to use artificial intelligence for business benefit. Inevitably, they look to IT for guidance. At last month's AI Summit in New York, Josh Sutton, global head of data and artificial intelligence at consultancy Publicis.Sapient, had advice for business and IT leaders alike around the basic ways that companies can and should start to work with AI.
AI can deliver value to companies by enabling them to derive greater insights from structured data, Sutton said. "We have a lot of structured data in our organizations today, in some cases a tremendous amount. When you start to layer machine learning on top of this, there's a huge opportunity."
In the advertising space, for instance, there's data from online traffic, TV, mobile devices and point-of-purchase sources that, together, compose a "terrifying amount" of information on roughly 90% of the population -- from what they buy to where they're physically located to what they're looking at online, he said.
With AI, companies can get valuable information by applying machine learning to that structured data. "That's something we could never do before because we simply couldn't scale to that level," Sutton said.
AI has already been applied to customer engagement in various forms, and there's an evolution underway.
First AI appeared in tools and appliances used for customer engagement.
More recently chatbots have emerged. "Chatbots are very good at taking natural language and extracting intent from that: 'Play the music for me. Turn the lights on. Tell me the weather,'" Sutton said. Following on chatbots will be digital assistants this year. "An assistant understands what you're saying from a contextual point of view within a narrow scope," Sutton said. Assistants can respond to a customer request with questions to clarify the problem it's working on and then make recommendations.
AI experts will follow assistants, in the 2018 to 2020 time frame, Sutton said, as assistants build upon their learning to develop a depth of knowledge within a specific domain.
Finally, savants -- those AI tools that are smarter than human beings -- will likely emerge, perhaps around 2030, he said. Companies shouldn't bother focusing on the last category, at least not yet, Sutton said: "That's fun for conversations but probably isn't as relevant as you think about your business strategy for tomorrow."
Accelerating business processes
Companies that are investigating how to use artificial intelligence should consider AI tools to automate or accelerate business processes, according to Sutton. But with today's AI tools, there's a significant investment required to reap the benefits: "For every day it takes you to teach a person something, you'll spend a hundred days to teach a technology platform to do that if it's an automatable task," he said.
For those tasks that are worthy of that investment, there are four main drivers.
- Cost reduction. The first, cost reduction, brings the lowest levels of return from AI, he said. Typically these returns come from automating manual processes and involve converting unstructured data to something that can be acted upon by an artificial intelligence tool.
- Competitive advantage. This driver, meanwhile, brings an important opportunity. "If you can do something fundamentally faster than your competitors, you can deliver better value to [customers]," he said. "If I can give you a commercial loan in two weeks as opposed to two months, that's a huge advantage for me going to market."
- Enhanced insight. The third driver is the enhanced insight derived from artificial intelligence, he said; a better understanding of the dynamics within a business can improve choices around, for example, product development.
- New business opportunities. Finally, the highest return from artificial intelligence will come from new business opportunities. "What are the opportunities that are out there that I could not engage in before?" he said.
More on artificial intelligence:
Special report on artificial intelligence: Applications, myths, development, data center, use cases, impact on jobs
How to use artificial intelligence in the enterprise