For all the giant leaps promised by artificial intelligence, when it comes to business, what we’ve seen so far amounts to just small steps. That’s not necessarily a bad thing. In fact, a number of very smart people advise companies to start small with AI: Use it to improve your customer service bots, for example, before you try to deploy it to cure cancer.
So, yes, that appears to be a sensible approach. But it can also be a dangerous trap. When you think small, notes this week’s guest, you get small results. Boston College professor Sam Ransbotham advises us that organizations also need to be focusing — right now — on what AI can do at scale when it is embedded as the organization’s central nervous system. In short, it will completely transform the way the company operates.
In this week’s episode, Ransbotham, who also leads MIT Sloan Management Review’s annual study on AI and business, focuses on these three big points when it comes to embedding AI in the organization:
- Start experimenting with it now — and it is, in fact, OK to start small.
- Take the initial gains AI provides in terms of efficiency and cost savings and invest them wisely — back into your people.
- Keep focusing on the next horizon.
For Further Reading
Sam Ransbotham: We spend a lot of time doing things we don’t necessarily need to — and a cool, shiny, new AI might let us do those processes better and faster — but it also might keep us from thinking about [how] maybe we shouldn’t do these things in the first place.
So I think about it like touching a hot stove. When you touch a hot stove, you yank your hand back immediately. You don’t convene a board meeting and think about it and discuss the properties of the stove and the relative heat. You pull your hand back.
I think that many of these leading organizations are stepping back and taking that step and saying, “Hey, we’ve got a little bit of breathing room here from the daily churn. Let’s see if we can push it to the next level.” And that’s different than just saying, “Hey, let’s, you know, knock off early and go home.”
Paul Michelman: I’m Paul Michelman, and this is MIT Sloan Management Review’s Three Big Points. In each episode, we take on one topic that leaders need to be on top of right now and leave you with three key takeaways for you and your organization.
For all the giant leaps promised by artificial intelligence, when it comes to business so far, what we’re really seeing are just small steps. Now, that’s not necessarily a bad thing. In fact, a number of very smart people advise companies to start small with AI: In other words, use it to improve your customer service bots before you try to cure cancer.
On the one hand, that’s a sensible approach. But it can also be a dangerous trap. When you think small, you get small results.
Sam Ransbotham: It’s relatively easy to have one-off AI efforts, and there’s plenty of cases where people can do one project or another project — save money here; save a little time there. But the question is, how do you make a more fundamental change? And that’s, I think, what many people and many organizations are struggling with right now.
Paul Michelman: That’s Boston College professor Sam Ransbotham. He leads an annual study with MIT Sloan Management Review and BCG that tracks corporate adoption of AI. This year, after surveying 2,900 executives across dozens of companies, the team saw an interesting and perhaps risky pattern emerging.
Let’s begin with what Ransbotham sees as the endgame for AI in the enterprise: an AI-based central nervous system.
Sam Ransbotham: An AI-based central nervous system is an organization that uses data and sensors to integrate all parts of the organization to work as a whole rather than a collection of parts. We tend to think about decision-making as being gathering data, everyone thinking about the data, and making a decision. But what AI offers is a different approach — an approach where the data can come in, and many routine decisions can be made instantaneously and automatically. And that’s terrifying when you think about your job [as] being making decisions. But it’s not terrifying if you think about your job [as] making a series of boring decisions or uninteresting decisions or routine decisions. And that’s, I think, the hope and the promise, and it’s sometimes the terror of this as well.
Paul Michelman: He uses the analogy of a child — or perhaps a daring adult — touching a hot stove.
Sam Ransbotham: When you touch a hot stove, you yank your hand back immediately. You don’t convene a board meeting and think about it and discuss the properties of the stove and the relative heat. You pull your hand back. And so what AI offers people, offers organizations, is the ability to react that quickly and to make decisions that quickly, and to ripple changes through an organization quickly. But that requires a more transformational change than just a series of one-off small projects.
Paul Michelman: Now, let’s back up here for a minute because this might seem a little too ambitious, at least for the next half-decade or so. As we noted a minute ago, companies are already seeing gains by using artificial intelligence to improve business processes and to save money. Isn’t it enough to just keep doing that — to keep finding more incremental improvements until AI is ready for its prime-time role? And, maybe even more important, until your company is ready for it?
Sam Ransbotham: If you just try to start and do everything at one time, it’s going to be a complete disaster. On the other hand, if you start small, you’re going to be stuck small. And that’s a classic sort of management problem. I mean, that’s what management is — balancing resources and making decisions like that. You have to find trade-offs. What’s a little different about AI is that it’s easy to get trapped in these too-small scenarios.
Paul Michelman: The reason small wins and incremental change aren’t enough, says Ransbotham, is that creating an AI-based central nervous system means completely redefining what a company is and what workers do — from the front lines all the way to the C-suite.
Sam Ransbotham: I think that’s where we’ve progressed enough with artificial intelligence to be able to see that we have another challenge beyond that. It’s like, you know, driving along and you see the first hill. Maybe we’ve come over that first hill, and we see the second hill, [and] maybe that’s a bit larger.
Paul Michelman: One of the big challenges is that AI at scale can’t be contained in traditional silos.
Sam Ransbotham: We’re used to having one group that markets, one group that does finance, one group that does accounting. And again, if those groups optimize within themselves, they become optimized parts but not necessarily an optimized whole. And the question — and I think we’re all trying to figure out this — is how much can AI offer here? Is it epsilon improvement over the individual optimizations, or is it a more radical potential for all working together, where boundaries are much more blurred — maybe boundaries don’t exist as we know them? These are just not very comfortable for us. They’re not comfortable for me.
Paul Michelman: Let’s bring this down to Earth for a minute or maybe even inside the Earth. We’re going to look at the example of the mining company Anglo American.
Sam Ransbotham: Their dream is that the spot markets in China tell their miners in Africa what to mine. What part of the organization are you going to scope that to? It’s tough to draw a line that defines just that group. It’s going to touch practically every part of the organization.
Paul Michelman: This kind of technology deployment — when it arrives in the not-so-distant future — will have all kinds of business implications that reach far beyond process improvement.
Sam Ransbotham: We spend a lot of time doing things we don’t necessarily need to — and a cool, shiny, new AI might let us do those processes better and faster — but it also might keep us from thinking about [how] maybe we shouldn’t do these things in the first place. My favorite example here is on health care and faxes. Lots of cool AI tools help you do optical character recognition, natural language processing — they can take a fax from a fax machine and image-process it and get that data into a new computer system. That’s great! But you know, that fax came from someone else’s computer system in the first place. So let me get this straight: We just took data from a computer system, created an image from it, put it across the phone lines, tried to reverse the image processing at the other end using cool AI, and put it back in another computer system. You know, I think it’s great you can do that sort of image processing, but why do it in the first place? Just skip the process. And so the danger is that AI becomes maybe more of a Band-Aid here. It’s helping you get past these current pains without helping you think [that] maybe these pains don’t need to exist in the first place.
Paul Michelman: And now we get to the existential problem: AI might take away jobs that humans are currently doing. Scratch “might.” It will replace jobs humans currently hold. It’s then up to us to figure out how humans can keep creating value — and maybe even more value.
Sam Ransbotham: If you think about what does AI offer, it can offer us all freedom to think more. It can free us from yet another boring email response that the machine ought to be able to do for you. It might save you from scheduling a meeting. But it might make you schedule meetings that shouldn’t exist in the first place. So what do you do with that free time? What do you do with that liberated time? And so I think that many of these leading organizations are stepping back and taking that step and saying, “Hey, we’ve got a little bit of breathing room here from the daily churn. Let’s see if we can push it to the next level.” And that’s different than just saying, “Hey, let’s, you know, knock off early and go home.”
Paul Michelman: The smartest future-facing businesses, Ransbotham argues, are already thinking about ways workers can be used differently with an AI central nervous system. And just as importantly, they are connecting today’s small improvements to the greater, transformative promise of AI.
Sam Ransbotham: So, what these organizations are doing is using these good processes. They’re using cost savings. They’re using time savings. They’re using process improvements. They’re using these things to create slack resources. When they have those slack resources, they don’t [say], “Quick, run and give everyone pink slips.” What they’re doing is thinking, “OK, how can we use what humans are inherently good at?” Being creative — thinking about the way things could be in a future that we haven’t yet seen — and trying to see how they can steer their organization to get there.
Paul Michelman: Ransbotham envisions a near future where the most successful companies are constantly ready to adapt. And that will be aided by both the capabilities of AI and the time that the technology frees up for workers to be more creative and thoughtful.
Sam Ransbotham: One of the things we’re trying to do here with this AI-based central nervous system is [to] be able to ... adjust to fit the changing environmental conditions. And we don’t need to think about this as a one-off/two-off thing. It’s about getting the organization working more cohesively and in a more unified organization.
Paul Michelman: That’s Sam Ransbotham, professor of information systems at Boston College.
All right, time to put a bow on it, gang: three big points about becoming a truly AI-driven organization.
Number one: Start experimenting now. It is, in fact, OK to begin small.
Sam Ransbotham: Thinking about those beginning points, thinking about starting small, thinking about getting some experience, save some money, get more efficient, gain that experience that’s necessary. This helps you get on that first summit.
Paul Michelman: Number two: Take those initial gains in efficiency and invest them wisely — back into your people.
Sam Ransbotham: When you get to the first summit, you will have slack resources. In theory, all those people, all that time you saved, all that money you saved will be available. What’ll it be available for? That’s a management decision now. What do you do with those slack resources? Summit the second hill. So the second hill is about thinking, “What’s next? What do you do with those slack resources?”
Paul Michelman: And number three: Keep focusing on the horizon.
Sam Ransbotham: So then, when you’re on the second summit, when you’re using those slack resources to improve your organization, then comes the third realization that there’s nothing but summits in the future — that the environment will continue to change, that your organizational constraints will continue to change, and that this is going to be an ongoing process that if you’ve got the skills built from the small starts, you can use what you’ve learned and how to use those slack resources to meet those oncoming challenges and changes coming forward.
Paul Michelman: That’s this week’s Three Big Points. You can find us on Apple Podcasts, Google Podcasts, Stitcher, Spotify, and wherever fine podcasts are streamed. We will be forever in your debt if you would take a moment to review and rate our program wherever you subscribe.
Three Big Points is produced by Mary Dooe. Music by Matt Reed. Marketing and audience development by Desiree Barry. Our coordinating producer is Mackenzie Wise.