Who Gets the Job of Figuring Out AI?

Aug 8, 2023 | Artificial Intelligence (AI)

As companies start to think about AI, the big question is who will manage this new job? Will it be technology chiefs or data officers or will new jobs be created for this position? Read on to learn more about who will manage AI.

As reported by Wall Street Journal on July 15th, 2023, by Chip Cutter.

The Big Question for Managers on AI: Who Gets the Job of Figuring It Out?

As companies race to incorporate artificial intelligence into their operations, many are wrestling with a more fundamental corporate concern: Who should manage such efforts?

The potential of AI to increase efficiency and change work in multiple white-collar functions—from marketing and human resources to finance and engineering—means it can be difficult to determine which teams should oversee the technology, executives say. Enabling too many departments to experiment with generative AI can expose a company to risks or data-privacy issues, executives say; restricting it can stifle innovation and potential productivity improvements.

“If you have too much top down, adoption really suffers,” said Hamid Moghadam, CEO of real-estate company and warehouse owner Prologis, which is working to find new ways to apply AI in its business. “And if it’s too bottoms up and unstructured, the good ideas don’t go anywhere.”

Across industries, CEOs have spent weeks debating the right ways to tackle AI projects, as the technology advances and companies attempt to seize on it. Some CEOs have asked technology chiefs or data officers to determine how AI should be employed. Such executives possess a technical background, an advantage in helping them assess which generative AI tools, such as ChatGPT, can be right for a business, said Alex Singla, a senior partner at McKinsey, who is advising companies on AI strategies.

Other employers are creating roles, like heads of generative AI, or forming committees to address the technology. Companies including VMware, DocuSign and others have pulled together small bands of employees from multiple departments to meet regularly to discuss AI efforts. These AI task forces, typically made up of about a dozen people, are becoming more common, corporate advisers say. Landing a spot on them has also become a coveted assignment for professionals hoping that involvement in AI efforts can bolster their careers, advisers say.

“Generative AI is the hot new thing,” said Paul Daugherty, chief executive of Accenture’s technology group. “People do want to become part of these initiatives. People want to get involved. They want to build their own skills around this.”

At VMware, the company created a centralized AI council earlier this year to monitor work across the business. It also started smaller AI task forces focused on specific departments, such as marketing, said Kit Colbert, the company’s chief technology officer. The central AI council meets biweekly and is made up of about 10 people, including representatives from the legal department, engineering, marketing and customer success.

One of the council’s first tasks: defining how AI could be used throughout the company. VMware approved the use of the technology for internal engineering projects, but not those involving products or software sent to customers, Colbert said. VMware agreed to allow generative AI in early stages of creating marketing content, but not in determining hiring or firing decisions in human resources, since the company wants to avoid potential bias in such actions, Colbert said.

The task force also helps to oversee internal AI experiments. To assess the impact of generative AI on its operations, VMware is conducting a controlled study within a group of engineers that builds internal tools. For a month, the company is measuring the productivity of about two dozen engineers using traditional software-development programs. Then, it will give those employees access to a large language model to help them in writing code using generative AI. It will measure whether AI makes employees more productive or satisfied, Colbert said.

“We’re going to be learning about: What does this team want? What do they need?” he said.

Many executives realize that AI efforts—and how a company is structured—will need to evolve. At MetTel, a telecommunications-services provider based in New York that employs roughly 750 people, the company is considering putting data scientists in every department to help teams find new uses for AI. Already, the company uses AI to help read emails from customers and to identify technical problems on customer networks before they are reported. CEO Marshall Aronow said he hopes it can do more within the company.

“This is a great opportunity to take away a lot of what we’ll call steady state or lower level or more manual work,” he said.

The enthusiasm around ChatGPT has led employees throughout the company to email MetTel’s chief technology officer with ideas about how AI could change existing processes.

Advisers say companies must be disciplined about what can be accomplished with AI now. Singla, the McKinsey partner who leads the firm’s AI division, said he recommends that companies determine two or three areas per department where generative AI could make a difference. Senior leaders or an AI task force can then evaluate which projects seem most doable and worth giving priority to based on the available data and technology. That avoids “a road map of 1,000 use cases,” he said.

Technology company DocuSign has assigned two executives to spearhead AI efforts: DocuSign’s chief technology officer, who looks after AI products offered to customers, along with its chief information officer, who is internally focused.

“We don’t want it to become 200 experiments in the company, which is why a limited number of people owning it helps,” said Shanthi Iyer, DocuSign’s chief information officer.

Some companies are trying to straddle a line between top-down control and bottom-up idea generation. The real-estate company Prologis is expanding employees’ access to a tool from Microsoft and OpenAI called Azure OpenAI. At the same time, a group of three leaders—the company’s chief technology officer, the head of its venture investing arm and a global head of strategy and research—help decide which AI projects are worth pursuing more broadly.

“We want to make it really easy for people to generate ideas and experiments. But then we want a level of control and due diligence on whether we can actually implement it,” said Moghadam, Prologis’s CEO.

Although some of AI’s potential seems like hype, Moghadam said, he expects it to change how some of his employees work. Prologis already uses the technology in pricing strategies and in finding efficiencies in some routine processes, like accounting. But it could prove to be helpful in determining where Prologis makes future investments, such as selecting locations to build electric-vehicle chargers near its facilities, he said. As models improve and companies turn AI loose on their own proprietary data, Moghadam said he predicts companies will get more value from it.

“It’s going to be kind of like the telephone and the internet,” he said. “I think it’s going to be embedded in everything that we do, and we’re not going to even think about it.”