This content is part of the Essential Guide: Special Report: Artificial intelligence apps come of age

Essential Guide

Browse Sections
News Stay informed about the latest enterprise technology news and product updates.

The future of AI apps will be delivery as a service

AI systems are generating huge hype right now, which makes it imperative for businesses to understand how the technology can be deployed most effectively.

Artificial intelligence has turned into one of this year's hottest trends in advanced analytics, but the technology being rolled out today may not be what many people imagine it to be.

Speakers at Dataversity's Smart Data Online said the future of AI apps is going to be intelligent applications delivered as a service. For example, a retailer may set up a customer service chatbot through IBM's Bluemix service built around the Watson analytics engine. Microsoft's Azure service is another example.

But this is a far cry from the sort of general intelligence many people think of when they hear the term artificial intelligence. Adrian Bowles, founder of Boston-based market research firm Storm Insights Inc., said general intelligence is still "light-years away."

Limited AI apps still offer value

At first this may come as a disappointment to those who imagined a future of conversing with intelligent machines capable of completing a variety of tasks. But upon closer inspection, there's still plenty of potential value behind the more limited AI apps being developed today.

Bowles said one of the chief benefits of today's AI apps is they will help businesses make sense of all the data they've been piling up since the early days of big data. AI systems will be able to quickly deliver insights from large data stores by using machine learning to pick out relevant information and natural language processing to write up human-readable reports, all in a matter of minutes.

"It's mostly about helping people discover what's in the data," he said. "It's about speeding up the process."

This type of AI still has limitations. Bowles pointed to the television advertisement for IBM Watson, in which the cognitive computing engine analyzes lyrics from Bob Dylan songs and then converses with the songwriter about the central themes of his songs. Bowles said the insights Watson delivers in this situation are at about the level of a college freshman. The analysis of lyrical content is technically correct, but it lacks a deeper understanding of the context of the times in which the music was released, which was a significant element of the music.

AI-as-a-service apps come of age

Still, businesses are likely to see a lot more of this sort of AI-as-a-service application in the coming months and years. Steve Ardire, a startup adviser at Kimera Systems, based in Beaverton, Ore., said he's seeing a "feeding frenzy" among software startups looking to release intelligent applications based on artificial intelligence. Chatbots are one of the biggest areas because they have potentially broad applicability, from personal assistants to customer service agents. Ardire said he sees this development happening in both consumer and enterprise software.

"You're starting to see the ramifications of this [trend]," he said. "Every application ultimately will become an intelligent application."

But all of this activity and interest in AI apps has inevitably created massive hype that enterprises need to evaluate in order to understand whether they have business problems that can be addressed by today's AI.

Hadley Reynolds, co-founder and managing director at the Cognitive Computing Consortium, said several leading companies like IBM and Amazon have marked AI as a major growth opportunity for the years ahead. This means marketing will only inflate hype even farther. In this kind of environment, it's buyer beware.

"The hype, both positive and negative, is all around us at this point -- far more so than even a year ago," he said.

Next Steps

Recent growth in AI is driven, in large part, by big data

Few AI apps represent the original vision for artificial intelligence

APIs and AI are a useful combo for making data more accessible

Dig Deeper on AI in the cloud