Videos
AI Helps Cut $4.5 Billion in Costs at IBM
Insights
- IBM's Client Zero program has removed $4.5 billion in annualized spend over three years, with another billion targeted this year, by applying AI and automation across the organization.
- Rebuilding processes around AI, rather than layering it onto existing workflows, produces the strongest gains in speed and creative output.
- Human judgment and craft remain central as AI absorbs derivative and repetitive tasks, freeing creative teams to focus on higher-value ideas.
What does it take to move AI adoption from scattered experiments to a fully rebuilt marketing operation?
Jonathan Adashek, Senior Vice President of Marketing and Communications at IBM, discusses how the company is reshaping its marketing function around AI, automation, and a sharper definition of what success should look like.
Jonathan outlines three critical shifts:
- How placing AI at the center of the function, then rebuilding processes around it, delivers deeper results than simply layering AI onto existing workflows
- How AI is compressing creative timelines and expanding the volume of ideas teams can explore, giving creatives more room to focus on original work
- Why judgment, skill-building, and a clear view of the ultimate goal - efficiency, cost, or better results - determine whether AI adoption pays off
Drawing from IBM's experience, Jonathan describes how the company built its Client Zero program to drive better results, greater efficiency, and meaningful cost reduction across every part of the business. He explains that rapid, ungoverned AI experimentation can drive token costs up quickly and leave organizations with disconnected pools of data that never inform one another. Jonathan argues that avoiding both problems requires a clear organizational goal before AI enters the workflow, along with the discipline to rebuild processes rather than simply automate what already exists. He also points to a shrinking shelf life for technical skills, now under three years, and stresses that companies need to give employees room to experiment, fail, and learn as tools evolve.
This interview was recorded at the Cannes Lions 2026 as part of a partnership between Infosys Aster and the Association of National Advertisers. Click to learn more about the ANA and the Global CMO Growth Council.
Dylan Cosper:
We are at Cannes Lions 2026, speaking with leaders driving innovation with AI in marketing. Today, I'm joined by Jonathan Adashek, Senior Vice President of Marketing and Communications at IBM.
Dylan Cosper:
Welcome, Jonathan.
Jonathan Adashek:
Great to be here. Thanks for having me.
Dylan Cosper:
Currently in your role, what are you seeing change within the marketing function?
Jonathan Adashek:
I think we have AI coming fast, budget pressure.
Dylan Cosper:
How is the marketing function changing with the rise of AI?
Jonathan Adashek:
I think the function is changing dramatically. A lot of people are saying, oh, we're going to put AI in, and they're looking at narrow scopes of it. I think the only way to really make the significant change that needs to happen for the function is to be looking across the entire operations of the function. Somebody described it to me earlier today as orchestration, art, and science, the three parts of marketing, and you've got to have it. It's not just in creative, because that's where a lot of people started a few years ago, and it's made a massive shift for us. Before we put AI into our function, our creatives' average amount of time on derivative assets was about 80%, and that doesn't allow them to be creatives. So how do you take that and really stretch that kind of engagement and those kinds of opportunities across the team and across the organization? But fundamentally, the most important thing everybody has to do is, what's the ultimate goal of what you're trying to do? You've got to have the ultimate company goal, the organization goal that's there, but what are you trying to do by putting AI in place? Are you trying to do it just to take out costs? Are you trying to do it to drive more efficiency? Are you trying to do it to deliver better results? What are you trying to do? For me, I want to make sure we're delivering better results. AI, automation, process evaluation, elimination, simplification. That... all that together should be delivering better results for the company, for our clients, for our partners, prospective partners.
Dylan Cosper:
What are some of those more worthwhile goals? You mentioned efficiency, you mentioned results. Can we get a little bit more specific on maybe some of the goals that you at IBM are pursuing?
Jonathan Adashek:
At IBM, we've looked at AI, and where we're putting it in place, and we've got a program called Client Zero. Client Zero is all about driving better results, more efficiency, and with it has come the cost takeout. We've taken $4.5 billion out of our annualized spend in three years. We'll take another billion out this year, and it's stretched across the entire organization. It's in every piece of the company. We're looking at it. We're taking a view of processes and what's going on. Sometimes it's AI, sometimes it's automation, sometimes it's just very simply take out a process that's been there forever, and it's only there because it's been there forever.
Dylan Cosper:
I guess there's also the aspect of when we're looking to move things to AI or automate, so to speak, there's a bit of let's not just create what we've already been doing, but now it's AI. It's probably a bit of redefining processes and changing the ways you work.
Jonathan Adashek:
That's exactly right. I think there's two ways to go about this. One is to take AI and just spread it across. One is to take AI and put it in the center and then rebuild around it. Re-evaluate how you do things, what you do, where you focus, how you use your resources. To me, that second one is the way to go, and that's what we're doing inside of IBM. AI is automating tasks, not necessarily roles, but the one thing that it isn't going to be able to take away is human craft.
Dylan Cosper:
Could you go a little bit deeper into that and really explain what you mean and why you think human craft isn't going to go away in the age of AI?
Jonathan Adashek:
I think AI is a voice. It's an input into topics. It's going to help take some roles, and some roles will change, some roles will go away, but with that, there's other new opportunities that will come, and we've seen that with technology over time in other instances. But if I look at AI, back to creatives for a second. AI is helping us create more ideas. It's taking ideation cycles down. We did something with The Sphere in Las Vegas. It took an ideation cycle down from what was projected to be fifteen to sixteen days, down to two days. And it gave more ideas to those involved so that they could become more creative. There's still that human touch that's going to be involved in a lot of things.
Dylan Cosper:
As AI adoption really starts to mature, what do you think is going to change? What are some of those top things you think will change within marketing? And beyond human craft, what's going to remain the same?
Jonathan Adashek:
Well, I think fundamentally what's going to change is how people are working and the knowledge people have to have. Because you've got to understand both how to use tools, but you've got to use tools, and you also have to understand how you work in an organization where your colleague might be using a tool that you're not on, and how does that change what's happening? Doesn't mean you have to be an expert in that. And I think giving people the skills that they need and the space to get those skills is really important. The average shelf life of a skill is, technical skill is just under three years now, so you've got to be constantly doing that. And I think it's going to be essential to create the space for people to feel comfortable, to experiment, to try new things, to fail, along with have great successes. I think that's going to be really important. And I think the importance of looking at how you hire people and what they're doing, and what you're looking for in those people is really important. One thing that is going to be very essential in this world going forward is just the simple fact of judgment. You're going to have to have a different level of judgment because there's fewer human touches on certain things.
Dylan Cosper:
What's the one thing in marketing around AI that isn't really being discussed but really needs to be?
Jonathan Adashek:
The cost. The cost. I don't think people are as focused on the cost of implementing AI in the organization if they don't do it right. There's going to be a cost. We all know that.
Dylan Cosper:
Absolutely.
Jonathan Adashek:
But if you just go out and say, everybody go create all the agents you want, then go, go, you know, have fun. A, there's a big token cost that can come with that.
Dylan Cosper:
Can balloon quickly.
Jonathan Adashek:
Yeah, very quickly. Very quickly. And then you're like, what did I get myself into? And B, you can then have so many different tools inside of an organization that you've got a pocket of data that sits on the right side over here and another pocket that sits on the left, and the right and the left never meet. That's a big problem.
Dylan Cosper:
If you have any final thoughts, if there's any key takeaways that you'd like to share with the audience.
Jonathan Adashek:
I'm just excited for what the future's going to hold. I'm thrilled to be with Infosys Aster. You all do some great stuff, and I was honored to be asked.
Dylan Cosper:
Thank you very much, Jonathan.
Jonathan Adashek:
Thank you.
