Zero One: Will You Pass the AI Challenge?
Business executives who bypass sluggish IT departments and buy technology directly from a cloud services provider, signing up for a subscription and flipping the proverbial switch, will find it won’t be so quick and easy with artificial intelligence, or AI.
Much like high-stakes poker, the AI game will test your patience, your mastery of data, the fortitude of your leaders, the culture of innovation (or lack thereof) at your company, and most importantly, your relationship with the CIO.
“Successfully implementing AI in learning software, with all its technical and ethical issues, will be a watershed event – separating businesses that can keep up with customers and those that fall behind and decline,” say Forrester analysts Brian Hopkins, Bobby Cameron, Ted Schadler and Rusty Warner in a research note.
AI, along with related machine learning and deep learning concepts, essentially means a computer has acquired the ability to learn how to do something, known as an “objective function,” through excessive testing on data streams.
The implication is that AI will be able to make better decisions and perform tasks more efficiently than humans, thus upending the relationship between man and machine – instead of computers assisting humans, humans will assist computers.
Related: Zero One: Can a Machine Acquire Human Intuition?
For business executives, the idea of a competitor getting to this place before you do should give you great pause. In fact, they might already be there. Thanks to maturing AI models and algorithms and cloud computing power to process massive data, companies are already getting real value from AI, says a McKinsey Global Institute report published in June.
McKinsey cites AI helping utilities forecast electricity demand, automakers create self-driving cars, and warehouses use smart robots. “In our survey, early AI adopters that combine strong digital capability with proactive strategies have higher profit margins and expect the performance gap with other firms to widen in the next three years,” the McKinsey report says.
According to Accenture, AI will bring the highest growth in gross value added, or GVA, to the following industries: information and communication, manufacturing, wholesale and retail, financial services, and healthcare. (GVA is a close approximation to gross domestic product.)
If you want to reap AI benefits today, however, you should have started years ago. AI has been likened to a child who learns through life experiences. It takes years for a child to become an adult, and what kind of adult emerges won’t be clear from the outset.
This way of thinking about technology adoption conflicts with business concepts of quarterly performance goals and quick returns on investment. AI will test your company’s stomach for innovation. For example, Zillow, which estimates home values, spent three years researching and developing deep learning without a business case to justify it, and is just now rolling deep learning out.
Zillow wants deep learning to be able to look at photos of homes and intuitively know which homes are more valuable than others. The research path headed by a chief analytics officer went from parsing photos to pattern recognition with imagery to deep learning to convolutional neural networks. (Yes, it’s tech intensive.)
Related: Zero One: Playing the AI Game
In order to be good at AI, you also have to be good at data – everything from gathering data from various silos of the company to securing data to labeling data for computers to crawl. It’s a complex and time-consuming process that requires data experts.
“AI systems will feed off fast data streams as rich sources for self-training models,” Forrester says. “You will need to build a system of insight with a data stream ingestion to connected instrumented software, machine algorithms, and self-training model management services in a closed loop.”
Forrester has a laundry list of important AI technologies: multimodal machine learning, machine learning servers, deep-learning platforms, transfer learning, emotion recognition, speech analytics, natural language generation, gesture analytics.
As a business executive, chances are you don’t know what any of this means.
There is no question AI development must be a team effort that includes the technology experts – that is, the CIO and IT department. Going it alone or just with the business side spells disaster.
For instance, at AI Summit in San Francisco in late September, a chief technologist at PwC told a tale of infighting over a machine-learning kiosk project at a retailer client. The kiosk was being developed by salespeople and a new chief digital officer without much involvement from the CIO and CTO. This resulted in the kiosk using the wrong product catalog and images, lacking plans to update data, and not connecting to back-end systems.
Clearly, there’s lots of deep technology expertise, data mastery, and patience needed to pull off AI. In a way, this can be summed up as another trial in the rocky, on-again-off-again relationship between business and IT. This time around, the stakes are at an all-time high with AI.
Can the aggressive, fail-fast nature of the business side accept the plodding, conservative nature of the tech side?
“It is a litmus test for firms of whether they can adopt joint business and technology organization collaboration models,” Forrester says. “Firms that recognize that AI strategy is business strategy – and elevate AI-savvy CIOs to the highest levels of business strategy input, influence, and accountability – will prosper.”
Tom Kaneshige writes the Zero One blog covering digital transformation, AI, marketing tech and the Internet of Things for line-of-business executives. He is based in Silicon Valley. You can reach him at tom.kaneshige@penton.com.
