As 2015 came to a close, the seventh episode of the Star Wars saga, “The Force Awakens,” stormed the world nearly 40 years after the first episode introduced us to characters ― some human, some droids ― that have become mainstays of our popular culture.
Our fascination with R2-D2, C-3PO and newcomer BB-8 confirms the attraction many people have to artificial intelligence and robotics. While 2016 is unlikely to provide robots that are able to sense, respond and adapt to people, this year should be a game changer in terms of the adoption of AI and the emergence of the Intelligent Enterprise.
The AI explosion
According to a new report from Tractica, the market for enterprise AI systems will increase from $202.5 million in 2015 to $11.1 billion by 2024. Enterprises will undergo significant transformations in 2016 as we see more applications powered by artificial intelligence, mainly due to advances in machine learning algorithms.
While the idea of machine learning is not new, the applications for machine learning are increasing due to the continuing explosion of data. Machine learning brings together a set of disciplines in computer science, probability, statistics and data science.
As machine learning algorithms become ever more evidence-based in their decision making abilities, their practicality in the enterprise increases as well. Fields such as speech recognition, natural language processing, surveillance, fraud detection, and robotics are likely to be some of the prime beneficiaries of improved machine learning applications in the year ahead.
Driving what Chris Bishop, managing director of Microsoft Research, predicts will be a breakout year for artificial intelligence is, “the emergence of new silicon architectures that are tuned to the intensive workloads of machine learning.” [These architectures should] “make virtual personal assistants like Microsoft’s Cortana and Apple’s Siri generally more helpful,” Bishop noted.
Cognitive computing hits the enterprise
Cognitive computing will also become more and more accessible to the enterprise. This development is being fueled by big investments from major players such as IBM, Microsoft, Apple and Google, as well as an array of startups pushing the boundaries of cognitive computing.
IBM Watson has been the poster child for cognitive computing since making headlines in 2011 by beating the top human contestants of all time on Jeopardy. Since then, not only has IBM made Watson 24 times faster and improved its performance by 2,400 percent, it has made it 90 percent smaller as well. To put those stats into perspective, what used to be the size of a master bedroom can now fit into the space of three pizza boxes.
Today, Watson is powered by 28 cloud-based APIs, a number expected to rise by 50 percent by the end of 2016. These APIs can build cognition into new applications, products and processes, leading to what IBM CEO Ginni Rometty predicts will be a future in which “every decision that mankind makes is going to be informed by a cognitive system like Watson.” Rometty has said she hopes Watson will create $10 billion in annual revenue within ten years.
Not to be left behind, Google’s $500 million acquisition of DeepMind launched its foray into deep learning applications. Google DeepMind’s initial focus has been on building powerful general purpose learning algorithms for mastering games like Go.
What’s more, Google’s recent announcement of the D-Wave 2X quantum computer ― which clocks in at 100 times faster than any computer today ― represents a way forward in tackling problems presently too hard or too complex for today’s computers.
Two global giants play catch-up
Even late bloomers like Microsoft are now in the AI race. Microsoft’s initial foray was Project Adam, whose goal was to enable software to visually recognize any object, no easy task given the immense neural network in the human brain that currently provides that capability. More recently, Project Oxford was launched to provide a set of speech, vision and language APIs.
Global chipmaker Intel recently acquired startup Saffron, a company that has developed a technology that ingests data from disparate sources and automatically connects the dots to further enterprise decision making. Based on a proprietary approach called Associative Memory which mimics the way the human brain draws conclusions from data and plots actions, Intel sees applications in real-time risk analysis, decision support in defense, detection of adverse events in energy, prevention of insurance fraud and improved medical diagnoses.
A new type of enterprise software
As AI systems increase in sophistication from “weak AI” applications focused on narrow tasks to more advanced “strong AI” applications characterized by sentience and applied intelligence, traditional enterprise software applications will undergo radical transformations as well.
These new “strong AI” applications will make use of machine learning algorithms designed to identify patterns by analyzing historical data to make predications. Applications that can mine data in real time and automatically discover insights and generate predictive models will become commonplace.
As “strong AI” applications develop, they will be able to do more than just support employees via traditional CRM, SFA and ERP systems. Some of these future AI systems may actually do parts of an employees’ job.
In fact, machine learning algorithms may do those jobs more precisely than humans by eliminating one traditional stumbling block inherent in enterprise software: the influence of human user input on data quality. Rather than undermining sales forecasting and CRM systems with inaccurate data and reliance on the constraints of relational databases, the “strong AI” frontier promises more accuracy and deeper insights.
Enterprises will also need to invest in new design capabilities such as Zero UI which emphasizes non-visual design. As AI powered applications become more mainstream, their interfaces will evolve to become voice, gesture and glance powered. Visual and touch screen experiences will be replaced by haptic, ambient and automated interfaces, some already evident in consumer devices such as Amazon Echo and Microsoft Kinect.
Cognitive computing will also hasten the demise of programmatic computing as computers will no longer rely on explicit coding to accomplish their tasks but rather learn by adapting to natural spoken or written language.
Intelligent customer service
Virtual call center agents will also be on the rise. AI will be applied to make self service faster and smarter, but will also launch its next generation of intelligent virtual agents. These agents will leverage existing IVR, CRM and PBX systems to handle customer interactions.
Gartner predicts that by 2020, 85 percent of customer relationships will be managed without human interaction. Initial experimentation and adoption of virtual assistants has already begun. In the US, for example, “Dom,” Domino’s Pizza’s digital assistant, is already guiding customers through the process of ordering pizzas using their smartphones. Dom has catapulted Domino’s into the public spotlight and positioned the company as a tech innovator.
Meanwhile, the Bank of Tokyo-Mitsubishi UFJ employs a customer service humanoid robot at its flagship Tokyo outlet. Standing just under two feet tall and weighing about 12 pounds, Nao robot works in the reception area fielding customer queries in Japanese, English and Chinese. Nao is not intended to replace branch workers; his mission is to free up branch staff time to spend on value-added customer service tasks.
While 2016 may not yet bring you a personal robot assistant, the year ahead is already shaping up to create a more responsive, intelligent enterprise AI environment.
The article was originally published on CMSWire on January 15, 2016 and is re-posted here by permission.