Companies focused on building sustainable AI in 2020

Fewer companies are deploying AI enterprise-wide in 2020 because of universal underestimation of investments and transformation required to have a rounded AI strategy.

Despite rampant AI development and coverage, expect 2020 to be a year in which companies take a step back from AI in many forms. The past few years of enterprise AI implementation has been targeted in siloes -- limited projects that are set to increase ROI or improve business projects.

The popularity of these platforms, algorithms or assistants has made headway across enterprises, but turning toward the new year shows that the phase in which companies take on limited AI projects seems to be giving way to more complex investments and this has led to a return to AI fundamentals.

Enterprise trust grows as deployment stalls

Companies that have been working with optimizing their processes through the usage of AI have reached an impasse. They have used AI on tasks of a repetitive nature, such as reading simple documents and transferring data from paperwork, and freed up employees to handle more advanced and important jobs.

"It becomes very difficult to read those things using existing technology and using AI substantially improved the ROI. So, that's why we say, using cool AI to solve boring problems," said Anand Rao, PwC's global artificial intelligence lead.

According to PwC's annual "AI Predictions" report, the number of executives deploying AI enterprise-wide within their businesses in 2020 is expected to drop significantly, with only 4% of executives planning to deploy AI enterprise-wide down from 20% last year.

Despite this decline, over 90% of executives surveyed believe that AI offers more opportunities than risks. Rao believes there is an explanation -- companies aren't abandoning AI rather they seem to be returning to AI fundamentals and refocusing on more sustainable AI. Learning the fundamentals and carefully explaining ROI around a topic that is continuously being redefined and reshaped has proved to be a difficult task.

The upcoming year is set to be one of more investigation into the thorough deployment of AI throughout the company. Part of this refocus is an emphasis on upskilling and cross-skilling. Employees who are specialists in one area should be relatively familiar with other aspects in order to better communicate and understand what is going on and what they are really asking.

"There's basically a set of three concentric circles: the entire circle of people in the company -- very conversant with AI, comfortable with AI. And then there is the middle group, which we call them bilingual; [they] understand the business, but they understand a little bit about data science to say what can be done, what cannot be done, and who is the right person to do it and then generally a small core group of people who are really data scientists," Rao said.

Getting everyone within the company familiar with the fundamentals of AI is difficult but necessary to create a long-term and sustainable AI strategy -- especially when business gains are often hard to prove. There must be a cultural buy-in by the employees for AI to continue to grow and this will take time.

Companies need to abandon siloes

Laetitia Cailleteau, UKI emerging technology and global lead for conversational AI at Accenture, points out that the companies that have focused on limited and separate deployment of AI are hindering themselves.

"We see a number of organizations that have a number of small projects really struggling to reap the benefits of AI," Cailleteau said.

Instead of infusing AI into the company as a whole, many executives have chosen to focus on individual projects. This has limited the benefits of AI and led to some underestimation on the difficulty of embracing AI as a full strategy. Now, as companies seek more benefits, they must take a step back and reevaluate what will be required to implement sustainable AI.

"People underestimate the transformation required to have a successful AI project within play. AI is a tsunami coming," Cailleteau said.

This tsunami is one that may be difficult to create buy-in for, however. With the move to more advanced usages of AI follows a need to convince boardrooms and employees of something difficult to explain and difficult to achieve, and executives must be able to acquire more finances, support and engagement as AI projects get bigger and stronger.

The internal culture of the organization can stall development. Applying an internal AI strategy requires transformative business changes and members of an organization are hesitant to dive in without seeing proof, Cailleteau said.

A journey to slower and steadier AI

The decrease in executives that plan to deploy AI enterprise-wide is also tied to the movement of companies to a more sustainable, and more transformative, deployment of AI. Implementing AI throughout the company means that department leaders are intertwining it within numerous processes to get little gains here and there; they must be able to prove that the investment is worth the time and money. Convincing all parts of the company to understand the plan and go along with it is essential but takes time.

"They need to have the right mindset. Change is never comfortable and the change required is actually quite big," Cailleteau said.

Rao believes that this step back is necessary when it comes to extending the technologies usage and sustainability. There must be research done by these executives on the best way in which to implement AI so that it is continuously working to benefit the company and not just a one-time return.

"Refocus on fundamentals and sustainability of what we are doing, as opposed to just a one-off win, a quick win," Rao said. "Those are all great to get the management buy-in and support for your project, but at the end of the day you need to show value on a sustained basis."

They need to start it off with the right mindset and approach as well. The demand for more ethical and explainable AI is unlikely to dampen and, as Cailleteau pointed out, consumers can be unforgiving.

"How are you going to be ethical and responsible and transparent?" Cailleteau said. "This is definitely a new function that people need to invest in when they do the transformation now, and they need to do it from the onset."

The AI market is booming, but it seems that most companies are pausing and reloading for more sustainable AI gains. Though it is a scale-back year for AI it still holds a tremendous amount of potential growth and billions of dollars soon. 2020 will be the step back needed for a more sustainable surge in years to come.

Dig Deeper on AI business strategies

Business Analytics
Data Management