Artificial intelligence is one of the buzziest terms in corporate America, but organizations don't need to pursue a tech upgrade as a solution for every problem, says Athina Kanioura.
Instead, Accenture's chief analytics officer says that companies — and the executives that run them — should meditate on the internal and external pressures that are forcing them to consider their investment.
Together, these pushes and pull form a firm's "AI narrative," says Kanioura, whose 3,000-data-scientist-strong Applied Intelligence arm at Accenture has consulted for organizations ranging from the Golden State Warriors to Carnival Cruises. A financial institution facing trouble growing its user base, for example, may opt to automate initial customer interactions to cut costs.
To vet your own AI narrative, you might want to consider the below factors:
Weigh the tech investments you've made so far and what return you are hoping to get by implementing applications that rely on AI.
For Kanioura, the first step in considering investing in AI is determining the appropriate return you are seeking and whether it will complement other tech priorities, like cloud computing or data analytics.
"Any investment in AI is incremental to existing investments," Kanioura said. "Every company should assess what investments they have made so far… and then see what is needed on top of that to drive a specific mindset change within the enterprise."
One limitation she urges firms to consider is the large upfront cost, particularly since profits will likely not come until after the first year, given most of that time is used to recoup the initial investment. Kanioura also advises companies to decide whether being an early adopter of the technology is critical.
While there can be value in capturing the "first mover" label — namely differentiating yourself from competitors — there are also potential pitfalls. A hasty switch to AI can disrupt global operations that may be centralized in one location but are used in many different markets, Kanioura said. Global corporations also tend to have very long business processes and are not as agile as smaller rivals, which can make implementation across the organization all-at-once more difficult.
Instead, she advises companies to identify areas of the business that are considered "no regret" moves, meaning there is room for experimentation without fear of any significant losses. One consumer brand that Kanioura worked with, for example, wanted to use AI to increase its market share in a specific category to 15%. It was fine with losing a portion of those sales initially in hopes of eventually reaching that goal.
Pick a section of the business to deploy AI that can provide easily testable results.
It's best to select areas that are more likely to produce tangible outcomes, Kanioura says.
Consider customer service: Companies like AT&T and LG Electronics use so-called conversational AI to address user inquiries through automated responses. Human agents can be looped in to the conversation when necessary, allowing employees to focus on more complex tasks. The hope is that the use of the technology will reduce response times and ultimately cut costs.
"You can get a lot of value by reducing costs, because now you deflect clients to the online channels and use that to fuel your growth," she said.
Another early-use case, according to Kanioura, is using AI to better target prospective customers, a tactic she argues produces clients that are more likely to purchase a specific product.
The technology, for example gives companies the ability to cater discounts to specific clients by quickly analyzing mounds of behavioral data, including what type of purchases they make and at what time of day. In one example that Accenture was not involved in, Best Western used information from an ad campaign that asked users to respond to specific questions to provide more catered travel recommendations.
Roll out the technology in a staged approach, and rely on both external and internal experts.
Automating operations like customer service centers will have a profound day-to-day impact. For one, it can displace the daily job requirements of many workers. To manage that, Kanioura said companies need to hire experts who guide corporations through periods of major change that understand the implications, as well as make sure they have in-house employees who are knowledgeable about the specific technology being implemented.
"That protects the company from massive losses in technology investments. It also protects the business stakeholders from making any decision that could be fatal," she told Business Insider.
Monitor results on a monthly basis, but aim to recoup the investments after one year.
Monthly costs are likely to change once artificial intelligence-backed services are introduced. More tailored advertising, for example, could gradually reduce the overall marketing spend by eliminating those users that are unlikely to purchase a given product.
"If you don't start getting results within the year," Kanioura said, "then there's a problem."
But the annual evaluation doesn't have to focus solely on direct financial improvements. Reducing how many individuals are focused on customer service issues, for example, can allow employees to pursue initiatives that in the long run result in revenue and profit boosts.
To Kanioura, the return on investment for artificial intelligence is twofold: achieving cost reductions through automation and creating new potential growth opportunities by freeing employees up to focus on higher-value initiatives.
Those benefits have the potential to grow significantly as AI matures and newer applications are able to replace more tasks currently done by human workers. It remains to be seen, however, how companies will respond to that from a workforce perspective. But it's likely Accenture — which is one of the world's largest firms consulting on the technology — will play a role in figuring that out.