January 9, 2019 | Written by: Wired Brand Lab
Categorized: AI for the Enterprise
Share this post:
Business leaders have gotten the memo: to be taken seriously, they need to take artificial intelligence seriously. Such business thought leaders know they need to give lip service to AI even if they don’t really understand it.
That explains why so few businesses are actively using AI right now. A 2018 study by MIT Sloan Management Review and Boston Consulting Group found just under 20% of companies were “pioneers” in AI adoption — which was the same figure as the year before.
Part of the reason for the inaction is that AI is still new and its ability to provide ROI is unproven, or at least assumed to be so. In lieu of blind faith, chief executives want to see concrete proof.
While AI’s promise is nowhere near fulfilled in 2019, many companies are using it to improve customer service, make better decisions and to squeeze efficiency out of their operations. Here’s a closer look at how businesses are applying these applications to their real-world operations:
Customer service: more than chatbots
24/7 communication has evolved from nice-to-have to a consumer expectation. 69% of consumers say they prefer chatbots for quick communications with brands. For instance, New Zealanders can receive instant information about health insurance by asking “Frankie.” Frankie is a virtual healthcare consultant from nib, the country’s second-largest health insurance company.
Frankie answers questions like “What does my health insurance cover?” and “How do I make a claim?” Thanks to natural language processing, it can recognize such questions in consumers’ own words and in Kiwi slang.
Like other businesses, nib is realizing one of the major benefits of virtual consultants — the ability to free up human customer service reps to take on more complex queries.
Ruchir Puri, CTO & Chief Architect, IBM Watson, said that customer service queries are either long-tail or short-tail. A short-tail query might be “I was logged in, but now I’m logged out, can you help?”
A long-tail query, meanwhile, might be a person calling an automaker. A consumer might describe how it was 20 degrees that morning, had trouble starting the car and saw white fumes coming out of the exhaust. To answer that question, the system would need to know some context, like the year and make of car, what the manual says and whether there are known issues related to that model.
“One of the clients stated to us that it typically cost them anywhere between $15 to $200 per customer service call, and even if we take the range of the lowest of that range which is $15, they were able to bring it down to slightly above $1, so roughly around 14X improvement efficiency,” said Puri.
Financial services and informed suggestions
In addition to understanding consumer queries in real time, businesses are using AI in financial services. Companies in the financial services industry use AI to make informed suggestions to consumers. For example, H&R Block is currently working with Watson to help consumers find tax deductions.
In that case, the company is using AI’s ability to digest large amounts of information. As H&R Block’s George Gaustello explained to The Verge, of the company’s 70,000 experts, perhaps 500 would know every deduction that a firefighter or farmer could get. To bridge that gap, Watson offers its own suggestions during the interview to help make the company’s experts smarter.
Squeezing efficiency from manufacturing
In manufacturing, the greatest value from AI can be created by using it for predictive maintenance (about $0.5-$0.7 trillion across the world’s businesses). AI’s ability to process massive amounts of data including audio and video means it can quickly identify anomalies to prevent breakdowns, whether that be an odd sound in an aircraft engine or a malfunction on an assembly line detected by a sensor.
Such predictive maintenance can help prevent poor practices that can reduce a plant’s capacity by 5-20%.
Uncertainty before taking the AI journey
According to Puri, generally the industries lagging the most in AI adoption are heavily regulated segments such as insurance, where a higher level of trust in the technology is needed. GDPR, the EU’s restriction on how companies can use personal data, is a major obstacle for these industries.
Fears AI might contain baked-in biases are another obstacle for these industries. The latter is the reason that IBM, for one, has launched a cloud-based service to find and remove instances of bias.
In general, the relative lack of investment in AI has more to do with a lack of knowledge about the technology’s ability to deliver real returns. Puri said the sooner that companies make the jump, the sooner they can see AI’s benefits and build on them.
AI OpenScale is the open platform for businesses to operationalize trusted AI and extend their deployments enterprise-wide.