Experts say the high failure rate of AI adoption is not a bug, but a feature: “Has anyone started riding a bike on the first try?”
Despite growing skepticism about artificial intelligence in the enterprise, three leaders from… MicrosoftBloomberg Beta and an AI startup came together in… luckLast week’s Most Powerful Women conference came with a unified message: High failure rates aren’t a bug in adopting AI — they’re a feature of learning how transformative technology actually works.
A panel discussion entitled “Working it out: How AI is transforming the office“, addressed the widely publicized Massachusetts Institute of Technology (MIT) study that suggests this Nearly 95% of enterprise AI pilots fail to succeed. This statistic has raised doubts about whether artificial intelligence can deliver on its promises, but the three committee members:Amy ColemanExecutive Vice President and Chief People Officer, Microsoft; Karen Kleinco-founder of Bloomberg Beta; and Jessica Wuco-founder and CEO of Sola– Strongly oppose the narrative that failure indicates fundamental problems with the technology.
“We’re in the first innings,” Klein said. “Of course, there will be a lot of experiments that don’t work. But has anyone ever started riding a bike on the first try? No. We get up, dust ourselves off, keep experimenting, and somehow we figure it out. It’s the same with artificial intelligence.”
Klein went further, encouraging the audience to become what she called “Vibe programmers“, or people who use accessible AI tools to build applications without traditional programming backgrounds. Coleman echoed Klein’s point, saying: “It’s all about the experience.”
She added: “We are on that winding border, where we will achieve some victories, and then we will see this rock bottom, and then we will achieve more victories.”
The Microsoft executive, who said her CEO challenged the senior leadership team to get the code up and running, emphasized that creating the right organizational culture is more important than the technology itself. “I think the study is really important because it actually reflects what a lot of people are feeling right now, which is: Will this really help me at work? Will it give me more happiness and less drudgery?” Coleman said.
Wu provided important context in trying to reframe MIT’s findings. “I think the actual study says that only 5% of the AI tools that people test get into production. What’s really interesting is if you take a step back and look at the percentage of studies of IT tools that were brought in that actually went into production before AI, they actually weren’t particularly high either,” she said, noting that success rates in deploying large enterprise technology were Historically around 10% or less.
Wu’s company, Sola, builds what it describes as “middle process automation” tools that help organizations automate manual back-office work. She stressed that the sheer volume of AI trials happening now makes low success rates inevitable. “I think there are a lot of tools that are being done right, there are a lot of tools that need to be tested, there are a lot of things that are being brought in,” she said. “At the same time, artificial intelligence is very new. It will make you hallucinate. You will have to work with experiences in ways that previous[generations]didn’t have.”
The conversation moved beyond defending failure rates to discussing what successful AI implementation actually requires. Coleman stressed the importance of building “AI fluency” across the workforce and recommended a collaborative approach where technical experts work alongside business users. “How do we bring together someone who’s really good at technology or continuous improvement, or some of these other types of advanced ways of looking at operations, and sit next to each other and not make something for you, but do something with you so they can learn how to put AI into your workflow,” she said.
Coleman also disputed the idea that enthusiasm for artificial intelligence devalues human labor. “The more we talk about artificial intelligence, the more people think we don’t trust humans,” she said. “It’s really important to talk about how important humans are in all of these workflows. So, it’s about talking about when I get the freedom to do what I can uniquely do as a human.”
Wu outlined what she sees in successful customer deployments: a combination of top-down leadership support and bottom-up engagement from employees who understand the daily workflow. “Leadership enables employees to test and build things that are obviously safe, but gives people the flexibility to experiment and try new tools, encourages them to use and build AI and helps them build fluency,” she said. “Your companies are full of people who live and breathe the work, and have been around for decades, sometimes even centuries. Hence, for AI to be deployed truly effectively, you need a tool to truly work alongside the people who do the work every day.”
Klein stressed that piloting does not require enterprise-wide deployments. “We also see startups working side by side, bringing engineers and business leaders together,” she said. “Even if we’re in a regulated industry, we can experience this in our personal lives and, you know, use it on the weekend to get non-sensitive information and just start to see some of how this technology works because that’s where you’re really going to get the wins and the advances and the big ideas.”
When an audience member asked what organizational conditions must be right for an AI transformation to be successful, Coleman’s answer revealed the cultural shift she believes is necessary. “You have to be okay with failure. You have to be okay with chaos,” she said. “We’re talking about the entry point into this transformation. You have to be okay with experimentation, and you have to be okay with that rough up and down.”
She added that companies need to embrace what she called a “learning organization” where “managers need to stop evaluating tasks and start teaching learning.” Key conditions include “vulnerability and courage” when organizations navigate technology that moves faster than previous shifts, she said.
The discussion highlighted a central tension facing companies: that the risk of moving too slowly in adopting AI may eventually outweigh the risk of experimentation itself.
You can watch the full discussion at luckThe strongest women’s event is below:
For this story, luck Use generative AI to help with the rough draft. An editor verified the accuracy of the information before publication.
Share this content:
Post Comment