Bonus Episode: Learn to Make the Most of Your Relationship With AI

**Allison** (0:03)
Hi, Allison here. Today and again in January, we'll be dropping a bonus episode in our feed. We're taking a winter break and we'll be back with new episodes in February. In the meantime, I hope you enjoy this excerpt from an MIT SMR BCG webinar based on our report, Expanding AI's Impact With Organizational Learning.
Sam and Shervin are joined by Walmart's Prakar Mahotra, who you might remember from our very first episode of Me, Myself, and AI. Prakar shares how Walmart thinks about five modes of human-AI interaction with specific examples. We've linked the related report in our show notes if you'd like to read more about our research on human-machine collaboration.

**Sam Ransbotham** (0:47)
Our main finding is that machine learning involves much more than teaching machines. Don't get tripped up by thinking that we just need to teach machines to do what we know how to do and that everything will work out.
Instead, organizations that are leading with AI focus on mutual learning, mutual learning between machines and humans.
So first, how did we actually do this? We're gonna talk about the research. What is the research? And well, what the research is is that we fielded a survey. We gathered more than 3,000 responses from managers and executives around the world. We then combined those survey responses with some detailed interviews with executives and academics across many industries. And the idea is that we can pull out a rich report that's both broad and deep. And we published this report. It's called Expanding AI's Impact With Organizational Learning. It describes those findings in details. We're focusing on organizations' impact, getting more impact, and finding that some organizations get more than others, and trying to figure out who does and who doesn't.

**Shervin Khodabandeh** (1:52)
We're seeing about 60% have some sort of an AI strategy. About 11% get some impact.
I'd also add a few statistics. 90% of organizations would say they believe that AI is critical and they have to figure out something to do with it, so there's clearly a desire. I mean, the major takeaway here is there is a big gap between ambition and impact, and so this report's really focusing on what do these 11% do that are getting significant in the order of tens and hundreds of millions for a typical multi-billion dollar company. And what these guys do is that they make humans and AI play nice together, right? They facilitate environments in the organization.
And by that, I mean full systems, right? For AI and humans to collaborate, to learn from each other in such a way that both the machines and also the humans get smarter.
They typically establish multiple ways for humans and machines to work together and learn from each other. And they do this in systematic and continuous ways. And this is really what we call organizational mutual learning with AI.
For this to happen, a lot of building blocks have to be built, right? So first, you've got to want to do something with AI, which is what we call the discovery phase, right? This is what, there is a strong desire to do something. There are limited deployments in very select and narrow applications. And there are some proofs of concept, minimum viable products. These organizations that are curious, have some activity, have some POCs. And what is the likelihood of impact is actually quite low, it's 2%, it's almost negligible. The next stage is what we call building AI. And this is where there are truly concerted efforts in the organization to build the foundation in data, in technology infrastructure, in digital and AI platforms, to hire the right talent and to create a cohesive business strategy that uses AI to drive real business value.
And actually that's where most of the lift's coming from. Now organizations that do these have a 20% chance of getting significant value from AI. Next step is scaling AI. Now this means going from the building blocks and the building of AI, from the production of AI to the consumption of AI. In other words, to implement AI in business processes, do it in multiple business processes. It also means not just looking at AI as a means of automation, but actually deploying it in revenue-generating functions in the organization.
Companies who do this double the likelihood of getting impact, so by this point they're at 40%. Now the last stage, which is really what we're gonna talk about mostly today, is about achieving organizational learning with AI, right? This is about building feedback loops, establishing these continuous ways where humans could learn from AI and AI can learn from human, knowledge sharing between humans and machines, and also it means allowing and facilitating multiple ways that humans and AI can work together. Organizations that do this have another 34% chance of getting value, right? So effectively what the research shows is that this is the secret sauce of these companies. So the major takeaway here is, first of all, significant impact is possible.