Q UA RT E R _ 0 2 _ 2 0 2 6 107 - - - opportunity, and it’s our role as leaders to help guide our people through it and use it to take our clients to a better place. EK: Even as they identify the opportunities that vol atility can present, I sometimes find CEOs struggle with feeling too constrained to take them—because they still have to hit the earnings and keep the train on the tracks. How do you juggle that? AK: One could go back and look at history. The pandemic came around, and everybody had uncer tainty about what revenues and profits would be. The debates were rife, “How bad is the dip? Is it going to be three months or three years? Will it be a 3 percent dip or a 30 percent dip?” You could say, “Well, I’m just going to try to manage that.” Or you can say, “I’ve got five other tough things to do. Why don’t I just package them all up and take it all on?” In our case, it was: OK, let’s borrow cash so that we are good through the pandemic. We’ve got to spin out a third of our revenue. We’ve got to divest some things that don’t fit our long-term strategy. And hey, let’s set a new M&A strategy. We used the oppor tunity to take on all the bumps in one go. And then, to try to grow. Deferring pain is never a good idea. EK: Let’s talk about one of the great opportunities: You are at the center of this moment around AI. What are some of the things you’re seeing in AI that are actually changing the way business works? AK: This moment on AI is so inter esting, because, in fact, the world has been using AI for about 30 years—there are many examples, but machine learning has long been used to make estimates in finance, in economics, and in sports. Then we had the era of AI that started with IBM winning Jeopardy! with a computer system called Watson. The problem with that era was that it was fragile and somewhat bespoke. You used a lot of data, and you used a lot of people to label the data; you constructed a model for a task, and if the task changed, or some of the data changed, you had to start again. What LLMs [large language models] now do is they put that on an industrial base. If more data comes along, you don’t need to start from scratch, you can just do more training—if the task changes only slightly, you don’t have to change anything at all. So, the moment it’s on an industrial scale, you can begin to deploy it. ‘We want to leverage AI to become the leanest, most nimble, and most productive company we can be.’ On the B2B side, I think the current genera tion of AI is going to help those who embrace it add about ten points to their bottom line while increasing revenue growth rates. Now, I’m not saying we would all add it to the bottom line. Rather, it frees up money for investment that you can plough back into innovation, R&D, sales. Take your pick depending on where you are on that journey. Where in the world have you seen an opportunity like that? At IBM, we want to leverage AI to become the leanest, most nimble, and most productive com pany we can be. I looked at a cost base of about $15 billion across what we would call third-party services, procurement, and G&A [general and administrative]. And I said, “We’ll take 20 percent out in the first two to three years, and then we’ll try to double that.” So, we set a bold target that we may not achieve, but if we get to 30 percent, I’ll claim victory. The idea is, this is not about tightening the belt and shaving 4 percent or 5 percent off; it’s reimagining how the work is done. I think almost every back-office function can become 50 percent automated using AI. We’ve done a lot in our HR processes; we have some thing like 94 percent of our basic internal HR transactions handled by an AI bot. We’ve also used AI at scale across about 8,000 of our 40,000 people who write code on products, and on that 8,000, we are getting 45 percent pro ductivity increases. Many people react with, “Oh, my gosh. They’re taking 4,000 people out.” But no, - - - - -