MIT, BCG research digs into the value of enterprise artificial intelligence: What employees get out of AI

Hype around artificial intelligence (AI) seems to focus almost exclusively on its benefits for the enterprise, but does its use come at the expense of the employees within the organization? Throughout the last 22 years, AI has replaced about 2.25 million manufacturing, and industry jobs — at the same time it is also expected to create around 97 million new jobs by 2025.

New research says no.  In fact, it does quite the opposite. According to a report from BCG Henderson Institute and MIT Sloan Management Review, 85% of employees say they directly benefit from AI in the workplace  — a stat that surprised the research authors themselves. 

“There actually isn’t much of a dichotomy, or forced choice between using AI to create value for the company versus the individual,” said Michael Chu, partner and associate director of data science at BCG, and co-author of the report. “What we saw overwhelmingly was the company benefits when the individual benefits… [It was] just so overwhelmingly positive in that way.”

Land O’ Lakes’ land of AI 

Chu told VentureBeat that the findings suggest that organizations that do have employees benefiting from their implementations of AI are up to six times as likely to see a financial benefit as well. 

For example, agriculture cooperative Land O’ Lakes. — which owns the buttery spread company by the same name — has been using AI to equip its co-op farmer members with the tools to make better, more informed decisions about soil and seed placement, soil nitrogen levels and even computer vision to spot threatening weeds or crop diseases, resulting in higher yield crop outputs. 

In the report, Land O’ Lakes CTO Teddy Bekel said that AI has benefited farmers: Rather than replacing their tools, it’s a supplement, he said, likening the farmer’s relationship with AI to “making decisions rather than following a recipe.”

The Land O’ Lakes example, the researchers say, showcases the benefits AI has for individuals, providing critical insights while still allowing their expertise and autonomy in their industry to shine, almost acting as a side-kick. 

Chu says it’s a perfect illustration of two major things that the report also highlights: A juxtaposition of what can be done with the hard sciences versus AI, and it provides a description of how companies can use AI to automate and assist, rather than replace.

“What’s new here is it incorporates AI providing recommendations, while not invalidating the farmers’ actual autonomy, so they can still decide what to do,” Chu said. “There’s a really great blend of the machine and the human decision-making on a micro level, as well as on a macro level.” 

Another key insight the example illustrates is that not only do organizations utilizing AI tend to benefit financially from doing so, but the implementation of AI can also help build relationships company-wide, strengthening culture. 

“In the Land O’ Lakes scenario, the farmers actually talk to one another about how to use the AI tools well,” he said, adding that for organizations, there is also a social element of AI that’s important. “One of the things we found, both this year and last year, was that often relationships between people inside the companies improve a lot because of AI. It’s actually quite interesting and useful in generating new ideas and new ways of working.”

Enterprises looking to derive the most value for their employees from AI should look into tools that automate workflow processes, act as a coach for employees to improve efficiency and that enhance customer experience, according to the report.

Where companies fail with AI roll-out to employees 

Of the many benefits organizations report from their implementation of AI, only about 8% say they are not experiencing positive gains.

This could be due to a couple of things, Chu says, ranging from lack of feedback from employee users, poor roll-out of AI tools or features and even design.

“In business AI, specifically, where you don’t have millions and millions of users but a smaller number of users, active feedback is really important for improving the product. If you’re not aware, and you’re not really providing feedback that can really make AI fail,” he said. “I think awareness, agency and trust are all pretty common areas of failure and in terms of rolling it out, failure can also be in the design itself.”

Failure to allow employees to have agency with the AI tools, Chu explained, typically comes from confusing pathways that lack clarity for users to have autonomy over the use of them. Similarly, lack of awareness, or not being educated on a tool or implementing one that then fades into the background without active participation, use or feedback also undermines an organization’s efforts to push for employee AI adoption. And, perhaps most importantly, a company looking to roll out useful AI tools should also create trust.

“Trust is not transparency, and transparency is not trust,” Chu said. “Trust is more intuitive in nature. Tell employees what it does. Introduce it to a team and tell them to just use it for the next few weeks. One of the things you don’t usually see when companies roll out AI is managers and senior leaders using it too. But having a senior leader use it who can say ‘I’ve used it, and it works,’ goes a long way’.”