I'm joined by Ryan McInerney, the CEO of Visa. Nice to have you.
And good to have a chance to talk. You know, being where we are, in particular in San Francisco, it seems appropriate to start with AI. Not to mention, on your most recent conference call, you spent a lot of time talking about the productivity gains that you've seen from it. I guess, big picture here in terms of the diffusion of AI at Visa, where do you think you are in that process, and has it gone more quickly than you might have anticipated or more slowly?
Well, it's great to be here. We're focused on AI broadly in two big areas at Visa. One is running the company better—more effective, more efficient, more productive, those types of things. And then the second is enabling agentic commerce and enabling all of us to have agents go out and shop and ultimately buy things on our behalf. On the first of the two things, it's actually developing more quickly than I expected. We're finding productivity gains across the company in product and engineering, but also broadly across the company that I think are developing even more quickly than I expected. On agentic commerce, maybe it's developing a little more slowly than I expected.
And I want to get to agentic commerce, and I know that has to do with trust issues, and you guys have been doing some work on that as well. Why, though, has it moved more quickly than you might have anticipated? I mean, I guess it's only been, what, not even four years since the introduction of ChatGPT. But I'm curious to see your own sort of thinking around that, what you're seeing within the organization. Because, Ryan, when I talk to a lot of CEOs, they talk about the friction of introducing a new technology into processes that have existed for a long period of time. So I'm curious to your experience.
We leaned into it very early. We made the tools broadly available across the company very early. We invested in a lot of user experience and deployment capabilities that made the tools easier for our employees to adopt. We invested in learning programs and hands-on-keyboard training programs, and I think that early investment has allowed our people to really put these tools to work and find productivity gains from them.
In our engineering teams, we're seeing 80% more code commits than we were seeing before. We're actually seeing an 80% reduction in the time it takes to prototype and build things that are ready to put to market in terms of feature development. We're actually seeing a 65% increase in the amount of features we're able to build at a quicker pace. So the results are encouraging. As a result, we're putting more product in the market.
Yeah. When you talk about those specific numbers which you shared on your conference call, I'm curious how broadly are those gains reproducible?
Yeah, they're very broadly reproducible. Those are the numbers we're seeing across a very large organization at Visa. And those are the product and engineering teams. We're seeing real productivity gains in other parts of the company as well, in marketing and HR and finance and risk and disputes and those types of things.
How much of the workforce then needs to be redesigned as a result of the things you're talking about?
I think in the big picture, everyone who's doing a job in somehow needs to think differently about how they're doing that job, leveraging these AI tools. I think that's true of every employee at Visa, which is why we've been investing so much in training and making sure people understand how to use these tools. Fundamentally, the workflows that almost anybody is doing at our company are being redesigned to be more productive, to be more effective, to be more efficient. And that means everybody at Visa is relearning how to do the best they can do in their roles. It also means we're taking people that used to work on some things in one part of the company and taking that role away, but then investing more in other parts of the company in other parts of the world to use these tools to drive more growth and more sales and more client engagement.
Right. And I ask that question in part because every time we see layoffs and or job reductions, you got some press around it. You guys didn't give a number, but you did give voice to it on your conference call talking about sort of tech and product teams and some reductions there. Some have reported about 2,600 jobs. We all want to say in the press, AI, AI, AI. Is that the case? Is that a fair assumption that AI is somehow at the root cause of why there might be reductions in force?
Well, for us at Visa, this has been a journey we've been on for a few years now. We've been driving more efficiency, more productivity, kind of relooking at how we're doing everything we're doing at the company. And as a result of that, there's been job eliminations. There were some we announced last quarter, there were some we announced a year before, and so on and so forth. So for us, AI is a piece of it, but it is really a journey around running the company better, trying to rethink everything we do to be as effective and efficient as we can, and ultimately in service of our clients, delivering the best products, the best services, and anything and everything we can do to help our clients be more successful and grow.
I'm curious as to how you think about it, though, broadly as a business leader, obviously running a very large organization with a very large market cap as well. There's a great deal of concern, let's call it, around what may happen as we become more and more reliant in these enterprises on AI. How do you see it in terms of its effect on employment? I'm talking more broadly here, not just your experience at Visa, but sort of throughout the economy.
I'm not sure I'm smart enough to know. You yourself and others have asked a lot of really smart, capable people that are seeing things across the spectrum, and they've got very different answers. That's the truth. And I'm just trying to do the best that we can do to run our company in service of our clients as effectively as we can.
So we'll just have to wait and see, I guess.
I think we'll have to wait and see.
It's obviously a debate we're going to continue to have. When it comes to agentic commerce, though, you're sort of at the front lines there. So what are you seeing? What tools are you providing, and what are the gating issues?
Yeah. So as I mentioned earlier, I think it's moving a little more slowly than we thought. But what we're finding is the web is having to be rebuilt as an agentic web. And there's a lot of infrastructure that's being built to enable sellers to make their inventories available to agents, to make their front door available to agents. If I had to summarize the work that needs to be done in one word, it's trust. We need to build capabilities and tools and ultimately have companies like Visa step in that can give consumers like you and I enough trust that our agent can go buy on our behalf with our financial information, and to give sellers trust that when an agent comes into their environment, it's a trusted agent. It's actually operating on behalf of David, and David has empowered that agent with a Visa credential to go shop at that retailer.
Yeah. I mean, given everything we hear about rogue agents at this point, I can understand why people might, at least at the very end, say, "I'm not going to trust an agent to actually do commerce on my behalf."
