The companies seeing real returns from artificial intelligence aren’t winning because they have better algorithms. Rather, they’ve figured out how to change the way they operate, according to a senior lecturer at the MIT Sloan School of Management and a digital fellow at the MIT Initiative on the Digital Economy.
Westerman has spent his career studying how organizations succeed at technology-driven transformation. In a presentation at the MIT Enterprise AI Forum in May, Westerman said that the companies succeeding at AI are asking better questions about their strategy and making a concerted effort to be agile.
“Technology changes quickly, but organizations change much more slowly,” he said. “For AI transformation, the hard part is not the AI. You’re not going to get any value from the technology unless you do business differently.”
Based on executive conversations and case studies he’s writing with companies such as Takeda Pharmaceuticals, global infrastructure company Ferros that business leaders should consider when implementing AI
What is your shared ambition?
Before deploying AI technology, leaders need to think about how they want to do business differently and why that would be better for their customers and employees. Without this framing, even well-re
Leaders at Rio Tinto spoke with employees about AI initiatives in terms of safety, not in terms of efficiency, Westerman said. They called automation a solution that will create “a mine where no miner will ever get hurt again.”
How will you govern your efforts?
Governance often functions as a set of rules designed to stop people from doing something wrong. But Westerman said that the companies making real AI progress have flipped that model by turning governance into a force that propels initiatives forward.
“Is your governance more the steering wheel or is it more the brakes?” he asked.
For example, a steering committee at HCA Healthcare examines the risks, business case, and feasibility of every AI use case the company employs (such as using AI to scan records or detect medical problems). It starts by asking whether an idea is big enough to create real value in the company’s hospitals and then poses a set of questions about risk and safety. The committee asks these questions before model development, again before the model is piloted at a small number of hospitals, and then again before the project starts to scale.
The steering committee also checks in periodically to see whether the models are still robust. Critically, the risk questions don’t stop progress; rather, they highlight areas that HCA needs to investigate as it makes progress.
Constantly measure performance, Westerman said. If project is not producing the anticipated results, change it or discontinue it. Governance should allow for speed, safety, and scalability.
How will you capture value by scaling AI efforts?
How can you make sure that your pilot project takes off across the organization? “We have this real problem right now that pilots are happening and then they’re not actually rolling out across companies,” Westerman said. “Different studies tell the same dismal story: Somewhere between 70% and 95% of AI pilots actually make it to scale in organizations.”
To scale successfully, think about how to use AI across the organization. For example, Dentsu Creative is using AI to help its people plan projects, develop graphics and copy, understand target markets, and even conduct campaigns.
Another point to consider is what might get in the way of scaling, including lack of executive support, which is one of the most common reasons projects fail. But lack of support isn’t always the fault of the executives.
“What are you doing to get that support before you start? And what are you doing to keep those leaders interested as you make progress?” Westerman asked.
Is your tech foundation solid?
Clean data, good tools, and integrated systems are fundamental to an AI strategy, but overlooking this step is all too common.
Westerman said that every organization he has worked with has had to do some level of data cleanup to implement AI solutions. Companies often try and ignore this problem or try to solve all of it at once, but neither approach works, he said. Instead, take it step by step.
Is your culture ready to move fast?
AI deployment requires new ways of working and a change to your organizational culture. Companies that are ready for the AI future have no problem experimenting and are willing to make decisions based on data rather than the way things have been done historically. Think in terms of speed, not perfection, Westerman advised.
DBS Bank, for example, engaged its whole workforce in learning about digital and AI topics, and in identifying ways to change the company for the better.
HCA Healthcare rejiggered the process of handing off patients between nurses at the end of every shift. A step that once took close to 50 minutes of intense cognitive effort now takes less than half that time and is easier because a computer helps with the script. Critical to the effort was engaging with nurses throughout the effort so they would be comfortable using the technology and confident that the model was performing as it should.
“These are big transformations,” Westerman said. “This is the whole company changing.”
Do your people have the right skills to work with AI?
No AI transformation succeeds if the people expected to carry it out feel threatened by it or aren’t familiar with the technology. Leaders should be clear with employees about how their roles will change with AI and what support they’ll receive so they can adapt successfully, Westerman said.
Many of the organizations he has worked with have proactively communicated the changes to their employees.
“They’re saying, ‘We’re going to change your job, but we’re going to help you make that transition,’” he said. “It’s really good to tell somebody, because if you’re not telling them, people are thinking the very worst.”
George Westermanis a senior lecturer in information technology at the MIT Sloan School of Management, a digital fellow at the MIT Initiative on the Digital Economy, and the founder of theGlobal Opportunity Forum. He helps executives understand the transformative potential of AI and other fast-moving digital technologies. His research studies on digital-ready culture and on workforce transformation provide important insights to move from transformation projects to transformation capability. He is also co-chair of the MIT Sloan CIO Leadership Award, which recognizes CIOs who lead their organizations to deliver exemplary levels of business value through the innovative use of IT.
For more infoSara BrownSenior News Editor and Writersbrown1@mit.edu
Ideas Made to MatterWhy AI-driven enterprises are the future of entrepreneurship
Ideas Made to MatterHow organizations can capture value from digital colleagues
Ideas Made to MatterAI financial advice can be good — especially with the right questions
