After a year of experimenting, testing, experimenting, and otherwise tinkering with generative AI, companies are now ready to start using it in the wild.
Earlier this month, a KPMG survey of 100 CEOs found that 39% expect to expand AI pilots company-wide this year. This number marks a significant change from November, when just 14% of business leaders said their companies were ready to use AI across their organizations, according to a Cisco survey.
However, implementing the world's most advanced technology company-wide comes with its own set of technical challenges. And even when that is resolved, companies will still need to train potentially anxious employees to use AI tools that some fear will cost them their jobs.
“It’s not about the technology; it’s about scaling,” Amy Challen, Shell’s global head of AI, said at the Fortune Brainstorming AI conference in London.
Companies of all sizes, from startups with a few initial employees to large Fortune 500 companies with employees around the world, are grappling with how to use AI. According to a January report from McKinsey, the companies most poised to succeed in the world of AI will not just be new innovators who have invented the technology, but existing companies who have successfully scaled it across their businesses. It will be.
For business leaders, the risk of being left behind is too much to accept, and as founders and technology executives discussed the topic on stage in London, businesses must choose the right uses for AI.1 One belief was clear.
A shotgun approach to using AI for everything won't work. To implement that targeted approach, IT teams can't be the only ones responsible for piloting AI. “The key for me has always been agile scaling with collaborative teams,” Challen says. “It's always a technical team, but it's also a business team. We never do data science without the business.”
At Shell, AI is already being used for what Challen calls predictive maintenance, allowing companies to predict when and where petroleum machinery will need repairs.
Sachin Dev Dugal, founder of startup Builder.ai, says the challenge for companies that know they will be leveraging AI in almost every aspect of their business is that employees are in the shadow of AI. The idea is to get used to the sensation.
For example, Builder.ai employees can have an AI assistant manage their calendars, and sales employees can give advice during calls with customers. It took some getting used to, especially because the idea of being listened to by an AI system was “very uncomfortable,” Dugal said.
“We need to reassure people that AI is making the calls,” he said during a panel discussion.
Panelists agreed that companies should not force AI on employees, especially if employees find it difficult to use. When companies find that new AI tools aren't being used company-wide, whatever their specific use case, Challen says they need to ask themselves why.
“If it's not being used, that's your problem, not the user's problem,” she said.
Employee hesitancy can be a major hurdle to implementing AI pilots company-wide. Generally, employees are willing to experiment with his AI, but doing so can have mixed results. Some people believe that it definitely improves productivity and gives them confidence going forward. Similarly, they are well aware that AI appears to be able to perform some of their jobs, which may deter employees from using the technology. A PwC survey of 54,000 employees found that 13% of respondents had such concerns. In fact, at least one study found that employees who use AI more are more fearful of it replacing them in the workforce.
But for now, the most common narrative about AI in the workplace is that it will become a co-pilot for workers' day-to-day responsibilities. Big tech companies like Microsoft, Google, and Salesforce have already begun marketing many of their AI tools as productivity tools, positioning them as the next evolution of word processors and spreadsheets. But now, employees are discovering that AI can do more than just help them with their jobs, it can do it all for them. And it presents its own learning curve.
“This is like an education to emotionally comfort people with the idea that they are one step away from how to perform certain tasks,” Dugal said.