Over the past several years, companies have shown a steady increase in their AI understanding, AI strategy and piloting of AI projects.
Despite this steady increase, however, more than 50% of companies report having some kind of AI working for them but only a fraction report financial benefits as a result, according to a new report by BCG and MIT Sloan Management Review. Why so few and how are the winners making it work?
The secret is learning; prioritizing organizational learning rather than just machine learning.
Organizational learning requires humans to learn from machines and machines to learn from humans. This requires fundamental rethinking of processes so that the combination of humans and machines is optimized and each is able to learn from the other.
Companies that combine AI projects with significant changes to business processes are 5 times more likely to realize significant financial benefits as those who make no or small changes.
This says that AI success isn’t just about the machines—it’s about the humans and the machines. In fact, in many ways, it’s more about the humans than the machines.
What does it mean to “learn from AI?”
First, learning between humans and machines needs to systematic and continuous. This means having feedback that allows humans to reflect on patterns that machines find. For instance, when a machine surfaces a correlation between a consistent behavioral trait exhibited by candidates and a specific prediction made by AI about the likelihood of success linked to the machine being taught what features are valued by humans in order to reduce patterns of systemic bias in a recruitment system.
Second, there should be many ways for humans and machines to interact. This means more points of contact between humans and machines. Designers should be thinking of ways to “embrace the seams” rather than create perfectly seamless experiences that remove all friction and therefore conceal learning opportunities.
Third, both machines and humans need to learn how to change. Processes and tasks can be reinvented with learning machines, which means that humans need to be able to imagine new ways to work. Companies that make extensive changes to many processes are 5 times more likely to gain significant financial benefits.
They don’t just change processes to use AI; they change processes in response to what they learn with AI.BCG MIT SMR
As the charts show, AI is much more than technical infrastructure, technical talent and an AI strategy. It’s really all about people and how interacting and working with a new kind of intelligence can unlock huge value for organizations who embrace the human side of AI.
At Sonder Studio, we help companies with activities that increase the likelihood of significant financial benefits by a combined 60%: strategy, designing solutions across use cases and—most importantly—sharing knowledge and structuring roles between humans and AI.