Early adopters of generative AI systems are already seeing return on investment for the applications they’ve added, Google Cloud found in a survey that was exclusively shared with Forbes. Google talked to 2,508 C-suite executives across the globe and found a variety of positive outcomes, with 74% already seeing some kind of ROI. Nearly half that implemented generative AI for productivity (45%) said it’s doubled. More than three-quarters (77%) said they have improved leads and customer acquisition as a result of generative AI. As far as revenue goes, 86% said they’ve received a boost of 6% or more.
How are these early adopters so successful? I talked to Oliver Parker, vice president for Global Generative AI Go-To-Market for Google Cloud about it. Parker’s position was created in January, and he works with executives on AI adoption every day. This conversation has been edited for length, continuity and clarity.
For you, what is the most important takeaway from the study?
Parker: [What] I think made the difference the most is at the C-level: This continued theme of C-suite sponsorship, executive sponsorship, C-suite commitment. You can say the same thing nine different ways, but the net is senior sponsorship and engagement on the use of this technology with an organization is incredibly apparent. Where you see success, that is one of the key ingredients.
The second one, I’ve been talking about since I started the role in early January, doing a lot of customer meetings. You’ve got all these use cases and all examples. It’s coming back and defining the use cases that are going to have the biggest impact—at least as an organization gets going—and the ones that have really solid ROI. What are the key metrics you’re going to measure the opportunity on? Is this for cost savings? Is this around new revenue growth? Really defining the business case that actually supports the implementation of the use case.
Oliver Parker, vice president of Global Generative AI Go-To-Market, Google Cloud.
©2024 All Rights Reserved
For me, what was very validating about this is success comes through deep adoption and commitment from the C-suite—which for me is not just exec sponsorship, but it represents a cultural shift that gets pushed down the org. And then the second one is the ROI that’s now starting to come out of some of these use cases going into production. Yes, we’re still early in the phases of this industry, but it is something that will absolutely endure over the longer term, and we’re starting to see some of the ingredients as to why.
How do you define C-suite buy-in and commitment? How long have these companies been committed to AI solutions?
When these tools started to show up—and ChatGPT was one of the ones that came out first, which really had that consumer feel in terms of the easiest way to interact with it. When people started to use these things to do their jobs on a daily basis, I think C-suites and all of us as consumers became very interested in it.
What I think really supercharged it was not just the value you could get from these large language models and the assistants that sit on top, but as importantly, almost the negative side of things, which becomes the security aspect of it. We all got very excited: Oh my God, look at all the things I can go do. All the things I can go find out.
The flip side is organizations and traditional enterprises are like: What information is being served up to these platforms? And how do I secure these kinds of capabilities so everybody can take advantage of them, but within the four walls of our organization? So I think two things happened. This experience was mind-blowing for many of us as consumers, realizing it could help our jobs. That combined with: We are an organization. We have confidential information we want to put into these systems. We want that to help. All this stuff that goes with governance and risk and data, combined with amazing sets of capabilities that are unlocking new things that could get done.
The accelerated engagement with someone that’s a C-level is exponential. Sometimes tech cycles might take two years. Right now, I think we’re working under such compressed cycles of how C-suites are getting involved. They see the value—whether it’s new revenue growth or cost savings—these kinds of capabilities, if introduced correctly, can deliver for the organization.
From what you’ve seen so far, what are businesses doing with AI, and where are they starting? Are they looking at will help the most? Is it more the easiest thing to do? Or are they looking at ROI as the first thing that they want to capture?
The hard response to your question is yes to all of them. You’ve got people that are very interested in it in their day-to-day roles. An enterprise where somebody at the working level says, ‘Look, these tools can help me do my work better, produce better quality in a shorter period of time.’ The CFO may be saying, ‘As I look at this stuff, I think there’s a huge amount of back-office automation that can take cost out of the business.’ Maybe we are using a third party to search through a whole bunch of documents for an insurance policy aggregation. We can do all of that through an LLM.
You’ve got different parts of the organization looking at [it], and I think it comes to life is where you’ve got a CEO saying, ‘This is a new technology paradigm that becomes a strategic differentiator.’ You get all those influence points. You’ve got a bottom-up culture, top-down support, and then in the middle, sets of skills and capability in engineering and IT that know how to harness this and create platforms in the environment where it works well for an enterprise, where it’s secure and all the things that go with responsible AI and security.
You need this confluence of activities, plus very strong support: This set of technologies will, if implemented correctly, become a differentiator for us as a business over time. And then [intent is reinforced] with cultural adoption: Go find experiments. Go build use cases. We’ll do hackathons.
The study showed 74% of enterprises using generative AI were receiving ROI already—and 86% were seeing revenue growth up 6% or more. Do you see that changing in 12 months?
For people that are putting these systems into production, that feels like the right kind of curve. What would happen going forward is the volume of people that would be doing it would be exponentially larger. I’d assert going forward that [those who already had] their first or second use case will see way more from their third, fourth, fifth, sixth, seventh, eighth to ninth.
It’s like, are you just having fun with this thing, or are you getting serious about this thing? I think there’s a phase in the middle. You’ve got experimentation, you’ve got production, and then you’ve got scale production: Rather than having one or two use cases that are really delivering value, you’ve got hundreds of use cases. I think this is what you’ll start to see. If we interviewed all those 2,500 a year from now, it would be fascinating to see the explosion of use cases that are now in production.
I still think that [74]% seeing ROI feels about right. Two things are going to happen. One is those that are using it already are going to be ahead, and it’s going to be a huge snowball with more use cases. Then you’re going to have a whole new bunch of cohorts jumping on. The question for the new ones is: Are the platforms better, where they can see a better ROI quicker, or are they going to follow the same trajectory as the ones that came in 12 months prior? My sense is it’s going to be somewhere in the middle, but it’s a massive amount more volume. Those that were using it are going to see exponential value, which becomes leapfrogging the new ones coming in.
According to the study, 84% of organizations with AI systems in place were able to move from pilot to production within six months. What is allowing this rapid adoption to occur?
The technology is so incredibly powerful. It’s not like it’s infrastructure, like when the internet was born and a lot of that buildout was happening. It was like, ‘This is going to be game changing,’ but it took a while before [anyone could] take advantage of it. This is incredibly obvious and valuable to the user very quickly. It’s such a huge leapfrog in capability of platforms to be this intelligent. Just the raw capability is leading to a different sort of adoption curve.
I think having a company culture that is espousing experimentation, that is asking for the workers to provide use cases where they think it can be valuable, that is doing hackathons, that’s number two.
There’s a lot of organizations, because of cloud computing being pretty pervasive, ones that we’ve been working with have been doing a lot of work around data: data transformations, rationalizing their data estate, and cleaning up data environments within their companies. That’s been a big initiative for many companies over the last three to five years. Having a lot of that data in a better position is also accelerating the adoption of generative AI. Where we see clients that have done big data transformations, we’ve seen incredible value accelerate out of generative AI because their data estate is also in order.
What are some of the more pressing reinvestment needs for the companies seeing ROI from AI systems?
The companies that have seen value from this, they will be directly investing back in talent. These are people with skills—whether that’s technical skills, marketing skills, [or] people that understand how to use these products in their day-to-day. They will also be investing back into work on their data platforms. I think they would’ve said ‘We’ve got our data in order for this one specific use case.’ They’ll see the value. They’re like, ‘Now let’s get the rest of our data estate in order.’
I think the snowball concept is very apparent. They’re saying, ‘We’ve got a bunch of ROI out of this. Now let’s supercharge it and move more towards this concept of scale production,’ where they take many more use cases into production. If you see the ROI, you’re going to reinvest, especially if you think you can get it from many other places across your organization.
Many of these customers [who responded to the survey] don’t have thousands of use cases in production. They’ve probably got a handful that deliver really strong value, but they know that over time they want to get to a thousand, because then you have AI at the core of how the whole company is run. As these systems get better and better, the value that’s inherited becomes more and more.
What wisdom would you share with a company in the 39% that doesn’t yet have any AI applications in production, but is working toward getting them in the future?
I think you have to do the due diligence [on] the way you think it can make a difference to your organization. Just firing something up to help you with emails could be helpful, but that may not be specific enough to show a true ROI. Really looking at the strategic imperatives of the business. With our clients, we go look at their quarterly earnings. What are the priorities of the client? Then we’ll sit down and work with them. If your priority is supply chain optimization, let’s talk about what your supply chain looks like. You have to line up to the business and the strategic imperatives of the client. That’s the first thing.
The second thing goes back to my earlier point. It’s not a surprise that the success and the ROI is coming from clients that have senior leadership support. It’s often a technology leader, but more and more I’m seeing leadership all the way up to the CEO, and also at the board level. Having senior leadership is absolutely instrumental. Lining it up to the business imperatives of the company is super important, because then when you sit down with the CFO and ask for investments in these new capabilities, you’re lining up to a strategic imperative of the company.
Thirdly, having the tech skills and the talent in-house is really important. Where I’ve seen companies do very well, they have some technical capability within their engineering or IT org.
And then finally, where it starts at the top, you have to infuse a culture of experimentation. You’re building a set of muscle and capability where people feel that they can experiment and be creative with these kinds of capabilities to help them. Some people will fail, but that’s the whole point. You have to go through this failure process to succeed at the same time.
