Here are the 3 steps companies that are just pivoting to AI should take to guarantee the process starts on a solid footing

Advertisement
Here are the 3 steps companies that are just pivoting to AI should take to guarantee the process starts on a solid footing

Simon Moss

Infosys

Simon Moss is the global head of AI and automation at consulting firm Infosys.

Advertisement
  • Companies can have a difficult time separating out the hype around AI with how the technology can actually benefit their operations.
  • The journey to adopt AI internally, however, is one that can take as long as 10 years. That's why Simon Moss, the global head of AI and automation at consulting firm Infosys, says companies need to start now.
  • The first step should be a small, low-risk move. But as the efforts scale, it's imperative that organizations form a central entity to manage all the AI projects, Moss suggests.
  • Click here for more BI Prime stories.

Interest in artificial intelligence is at a fever pitch, but it can be difficult for corporations to determine whether it's all hype or if the advanced tech can actually improve operations.

While there remains a healthy amount of cynicism around the technology, it's imperative that organizations begin to think about incorporating it now - especially because the journey can take as long as 10 years, according to Simon Moss, the global head of AI and automation at consulting firm Infosys.

"The decisions around a cluster of separate but deeply related technology innovations are existential to whether a firm will be strategically successful or not over the next decade," he told Business Insider. "It's a huge challenge. Not a wave, but a tide as impactful as the internet, transforming the very DNA of the enterprise."

Companies are already using AI in a number of different ways, including to help figure out whether store shelves need to be restocked or to cut back the number of potential applications that human resources departments must review for open positions.

Advertisement

Still, the vast amount of AI efforts fail. That means to avoid the pitfalls, executives need to focus more heavily on answering a key question: how do I get started?

Infosys is an organization with over 228,000 employees that has helped companies like Pfizer and e-commerce firm Radial, Inc. on their digital journey. At the firm, Moss has been spending ample time helping companies overcome that challenge.

He shared the three steps companies should take to bypass the hype and start their AI journey off on solid-footing.

Determine the end goal before starting

What companies are hoping to get out of AI is likely to not only vary by organization, but also within various business units. That can make figuring out the overarching goals more difficult.

For some, it may be achieving operational efficiency in supply chains. For others, like retail companies, it can be better understanding the end consumer to enhance marketing efforts and tailor promotional offers.

Advertisement

It's a critical part of the foundation of AI-based projects because it helps inform other questions, including the state of the digital infrastructure and whether the initiative can succeed. Without the right stored data, for example, it can be impossible to power the applications.

Begin with 'very small projects' that can be easily measured

Gone are the days of sweeping, enterprise-wide projects that can cost upwards of hundreds of millions of dollars. Instead, like other consultants, Moss advises companies to start small.

Organizations, for example, could relatively easily develop and use technology like natural language processing to understand, organize, and consume vast amounts of documents. Or they could employ robotic process automation (RPA) in sectors like customer service that can, among other things, help agents quickly compile information on those calling in.

It's a common area for companies to start their AI journey on. In fact, Gartner has estimated that spending on RPA could top $2.4 billion in the next three years. The key, according to Moss, is choosing a low-risk area that can produce a quick return on investment.

"That gives a client a sense of confidence they can control their destiny, and frankly control their careers, that they're not signing up for a $100 million system-integration project," he said.

Advertisement

As employees become more comfortable with the technology, Moss says executives can begin scaling those initially rudimentary projects to incorporate deeper machine learning and drive a higher value.

Create an AI 'center of excellence'

Often, the various business units within a company will be testing out different applications.

Human resources, for example, could be using AI-based tools to cut down on the number of applications they review per open position. Meanwhile, store operations might be tapping into the technology to try to reduce shoplifting incidents.

That's why having a central entity to manage all those projects is critical.

Organizations need a "marketplace that allows a common philosophy, that allows common security - particularly around personal identity information - that puts in a framework of scale and resiliency around design, and begins to impose a common philosophy around what is a huge amount of intrepreneurial endeavors," Moss said.

Advertisement

One major benefit is to cut down on the number of bespoke efforts. Instead of different departments hiring their own data scientists to pursue AI-based endeavors, a so-called "center of excellence" can flag any duplication and find ways for platforms to be used across business units.

Read more: The head of IBM's Watson walks us through the exact model tech leaders can use to build excitement around any AI project

The involvement of such an entity, however, will vary by organization. Those just starting out on their journey, for example, might just use it as a project database to keep other leaders informed. More established companies might employ a center to ensure all initiatives align with a core philosophy that define and guide all AI-based efforts.

One thing to avoid is placing too many protocols around how the entity operates, argues Moss.

"You don't want to suffocate that with too much bureaucracy and administration," he said. "That's a balance between discipline, rigor, and innovation that is something that cannot be designed as a cookie cutter [model]. It has to be specific to the culture of the customer."

Advertisement

Bypassing the hype and starting on an AI journey can be difficult. But these three steps can serve as a key launchpad to help organizations at least start to formulate a strategy.

{{}}