Reset and reimagine: The role of generative AI in SG&A success

In this episode of McKinsey Talks Operations, recorded at McKinsey’s recent CFO forum in partnership with Inside the Strategy Room, host Daphne Luchtenberg talks with Matt Jochim and Martin Rosendahl, partners in McKinsey’s London office, along with Samdruk Thomas, an associate partner based in Zurich. Together, they delve into the challenges and opportunities faced by sales, general, and administrative (SG&A) functions in today’s fast-moving economy. Discover how courageous leadership by CFOs can balance short-term cost pressures with long-term innovation, leveraging digital technologies such as generative AI, and how a clear strategic road map can deliver a successful transformation.

Their conversation has been lightly edited for clarity.

Daphne Luchtenberg: Your company’s future success demands agile, flexible, and resilient operations. I’m your host, Daphne Luchtenberg, and you’re listening to McKinsey Talks Operations, a podcast where the world’s C-suite leaders and McKinsey experts cut through the noise and uncover how to create a new operational reality.

Sales, general, and administrative costs, sometimes known as SG&A, G&A, or back-office costs, are under pressure. These overheads are facing constant scrutiny as companies look to manage costs in a fast-moving and often volatile economy. Today, we’re bringing together three experts from our Operations Practice to talk about the pressures leaders of these functions are facing and the role of digital technologies, including generative AI, in helping them to succeed in the future. They’ll be talking about opportunities to reset and reimagine SG&A functions, a topic covered in our recently published compendium that you can learn more about if you visit our podcast page on McKinsey.com. Let’s get the conversation started.

Welcome to Matt Jochim and Martin Rosendahl, partners in our London office. And Samdruk Thomas, an associate partner based in Zurich. Matt, let me start with you. What are some of the fundamental pressures that G&A functions are facing today?

Matt Jochim: It’s interesting that we saw clients and companies go through COVID and really lean into making some tough choices and navigating uncertainty. What we saw last year is that companies seem to have acknowledged the need to continue to do more—to be competitive and focused on the right investments. And what our research says is that most companies set ambitious targets for themselves, but few actually delivered on the targets that they had set. Now that we are in a world where we’ve got increasing complexity, geopolitically increasing pressures in terms of commodity inflation, and labor uncertainty, and all these things going on, I think leaders are having to say, “What’s next?” We need to face this and find ways to be more productive, but the answers aren’t easy or straightforward.

Daphne Luchtenberg: Thanks, Matt. And Martin, are there other challenges that leaders in this area are facing? What advice are you giving your clients?

Martin Rosendahl: What CFOs ask us right now is, how should I balance the short-term cost pressures driven by inflation versus the more long-term innovation of these functions? They are asking about AI, generative AI, and process optimization. So what we’re advising clients is that this is a time for courageous leadership by CFOs. It’s a time for CFOs to step up and reset their functions but also to reimagine. And courageous leadership will display both at the same time. So optimize your cost base short term while innovating for the future through AI and technology.

Matt Jochim: And if I can build on Martin’s comments, these sorts of exercises oftentimes aren’t fun. Like the reset piece of this: the employee base has broadly gone through restructuring and belt-tightening, and that sort of thing. But it’s a necessary part of running a healthy P&L and continually reallocating the company’s resources, both financial and human resources, to be focused on the right things. The reimagine side of this is where the inspiration can really come from in terms of helping improve the competitiveness of the business, giving colleagues the tools that they need to do a better job—really creating distance between yourself and your competition by finding new ways to serve your customers more effectively. So I think that this is beyond just the idea of how you solve the financial puzzle. There’s also some real opportunity here to lead, motivate, and inspire the organization to do things better, to be a healthier business on the other side.

Daphne Luchtenberg: Samdruk, I’d love to bring you in here to talk more about the reset and reimagine approach. How is the reimagine phase showing up in reality?

Samdruk Thomas: When it comes to reimagine, one of the challenges that we’re seeing is that there is a certain disconnect with CXOs [C-suite leaders]. It’s a disconnect between the level of aspiration where we see that the majority—over two-thirds of CXOs and companies—are investing significantly in digital and analytics, but at the same time fewer than 25 percent, based on our research, are actually seeing the benefits and managing to capture sizeable value from these investments.

Daphne Luchtenberg: Thanks, Samdruk. And I’d like to stay with you for a while as we switch our focus slightly to digitization of G&A functions, a key element of the reimagine phase. How are you seeing companies use digital to enable G&A transformations?

Samdruk Thomas: When we look at digital, there are two waves of digitization in G&A and support functions. The first wave has been more focused on robotic-process automation and what we call analytical AI—leveraging AI to solve problems faster, more efficiently. But what we now increasingly see is that the second wave, which is building on generative AI and leveraging AI to create new content, generate new insights, is now having more widespread adoption.

Martin Rosendahl: And the impact on functions like finance over the next three to five years will be very significant. Some clients that we work with have found that there is a rather simple recipe to think about: “How should my workforce shift with the increased importance of technology?” And they would typically look at, first, which roles are augmented. For example, in financial forecasting, I’m a financial planner, I’m an analyst. What does the intelligent forecast tell me? Second, there could be roles that are accelerated. I can do my job quicker, I can close quicker. Then there are jobs that will literally get automated. We talked about this for some time; this wave is now really happening. And then there’s a fourth category, typically forgotten, which is added roles—the new roles that we don’t have today in the finance function that will be critical to get the full benefit of AI and generative AI. So we have the augmented, accelerated, automated, and added.

Matt Jochim: If I can build on that, we see a lot of conversation about things like generative AI: What does this mean? What’s the buzz? Is this buzz? Or is this real? I think our view is, on the one hand, that we don’t fully know how this is going to play out, just like we didn’t know how wireless technology or the internet was going to enable business. But we know that this is big, and this is a real shift. So what we’re advising clients to do is to familiarize themselves. And we think that in many businesses, it makes sense to begin making some investments to start to explore and understand the potential of how these sorts of technologies could add value. But it is going to take a little bit of time for this to sort itself out and clarify what’s the real at-scale business value.

Martin Rosendahl: That’s a great point, Matt. And sometimes a CFO asks us, who is the best in AI? Where should I look for inspiration? Am I ahead or behind my competitors? And in one way, it is not really the right question to ask. We’re in the early days here. This is a marathon; we’re in the first or second kilometer of this. So if you’re a couple of meters ahead or behind, it doesn’t really matter. What matters is, do you have a strategy? Are you building the right capabilities? Are you testing and trying fast? Are you having the right partnerships? Are you open for innovation? Those are really the questions that CFOs should ask right now, rather than, am I slightly ahead or behind my peers?

Samdruk Thomas: Yes, and I think there are three critical success factors that we advise CFOs and CXOs to take into consideration as they embark on their digital journey, and with generative AI. The first one is more on the technical part—having the right solution architecture in place. But that is actually only the foundation. What is even more important is having a business spec transformation approach: identifying the right domains and use cases that will deliver value. So it’s less about, “How can I apply generative AI?” and more about, “Where will it actually deliver benefits to the company?” And the third factor that Martin already mentioned is around people and change management—building up internal capabilities.

Matt Jochim: And I think this last point around people and capabilities is an easy one to overlook. It’s easy to get focused on processes and technology. What we’re seeing in many companies is that the talent and people become the bottleneck on how you are building skills and culture.

Martin Rosendahl: Here’s an example from one of my clients I worked with for many years. They had literally invested in every technology that exists. They had acquired software; they had all cloud solutions. Anything that was out there, they had it. But the value wasn’t coming through. So when we started looking under the hood, [we saw that] it was really about having the right capabilities for them to make use of all of those exciting and fantastic technologies. And I think that’s really a core lesson that we see in general with clients. They do have a lot of technology, but do they have the right capabilities to make the most out of it?

Daphne Luchtenberg: Thanks, Martin. That’s really interesting. Moving on, I’d like to get a bit more granular for a moment on one element of digitization—data quality—which seems to be a pretty big stumbling block for companies. What’s your experience with clients here? Over to you, Matt.

Matt Jochim: One of the things that we often hear from clients is, “My data is poor, my data is too bad for me to do some of the things that you’re describing.” Counterintuitively, what we see in most companies is that the whole idea of putting the data to use is what improves the data. If the data is just sitting in a back room, nobody really understands it. When people don’t understand what it’s used for and why it matters, it is difficult to improve it just as a corporate enterprise exercise. I’ve seen plenty who have failed trying to do this. With a master data optimization effort, what we tend to see is that it is only once the data is being used, that people say, “Wait a minute, that data is wrong—we need that fixed to make the right decision.” Then the organization kicks into [a new phase of], “OK, maintaining and improving the data is now what we do to run the business.” So a bit counterintuitively, stepping into applying and using the data is one of the critical enablers to getting data of sufficient quality to get good analytics and outcomes.

Martin Rosendahl: Many clients in this situation also hit some boundaries. And clearly, with increased cyberthreats, many clients have put boundaries on how to use and access data. So in the approach that Matt was describing, it is helpful to work very closely with your cyber team, with your compliance teams, and make sure you find the right balance between protecting the enterprise and protecting the organization, making sure that the company can make use of all of the internal and external data that they have available.

Matt Jochim: Let me give you an example. A couple of years ago, there was a really big push around zero-based budgeting, and companies were going through the process of really applying costs at the general ledger code and cost-center level. And inevitably, when you would go through and look at the historical spend, there would be many, many misclassified POs, when in reality, the company didn’t care because they hadn’t been using that data to actually drive any decision making. Now, you pivot towards the spend being really scrutinized at a very granular level, and it starts raising questions like, “Wait a minute, you said that it’s IBM for business software. But it’s actually consulting, so we should be classifying that as consulting.” And so, working through that process led to the recognition of why classifying spend data correctly matters. And that leads to accountability and the actions to maintain the data in a healthier, higher-quality way.

Daphne Luchtenberg: Thanks, Matt. And of course, it is important to look at the challenges, but what about the opportunities? Martin, can you give us a vision statement on what you think digital solutions can do for G&A functions in the next three to five years?

Martin Rosendahl: That’s a fantastic question. What is possible today is around three areas. It is possible today to allow access to data—financial, operational, and commercial—to the broader organization to make the right decisions. It is possible today to automate the most basic jobs and drive value out of that. And it’s definitely possible today to have driven forecasts in finance or in procurement or in HR. What is possible in three to five years, we can only speculate. But our view is that it will look significantly different from today. So again, we are back to having the muscle to innovate, to be fast, and really follow this evolution of technology from a business value perspective.

Matt Jochim: Maybe if I can add two more examples. Example number one: I think that when it comes to planning and coordination of activities across functions, in many organizations, the data exists in different pockets, but they’re not using common data sets, common assumptions to drive cross-functional decision making. That is available today, and there are several technology solutions that can help enable that. Another enabler that’s available today that we’re seeing companies step into, that’s a real game changer, is using these sorts of digital exercises to break out of functional silos.

Rather than think this is what HR does, and this is what finance does, let’s think about end-to-end processes, like procure to pay or hire to retire, and then just start to think about why we need to hand this off between functions. Why do multiple approvals need to take place? Why is this information being entered in multiple places by different colleagues? Just taking that end-to-end view can not only unlock digital opportunities, it can also change processes, the ways of interaction, and speed up decision making to streamline the organization. So I think one of the big learnings of all this is doing it holistically, where the digital and technology can be the unlock, the breakthrough, the shift, or the catalyst for this sort of reimagined effort. These are things that can be done today.

Daphne Luchtenberg: So you’re saying that technology is becoming a catalyst for reimagining G&A functions. But what about one of the topics of the moment—generative AI? Is it table stakes or is there a real lasting competitive advantage to be had by getting involved now?

Martin Rosendahl: The way I would probably frame the question is, “Is it here and now driving business value, or is it still in early stage?” We see that some of the underlying technologies have existed for years. We see organizations driving the value of that to fully make use of some of the more consumer type of solutions that exist, but that is not being done at scale in most organizations today. So these are still areas where organizations are exploring. It is an interesting dynamic; we see many teams challenging their organizations on these topics. Why? Well, because they are speaking to investors, they’re at conferences and organizations, and they’re getting a lot of questions on how their organizations are taking advantage of generative AI. They go back to their organizations, and they are facing five, ten, 20 big issues with how to do that in the corporate world. So here is where the executives really need to unlock the internal boundaries that have existed to make use of this type of technology.

Samdruk Thomas: And with generative AI at the moment, it’s a very fast-paced field, and it’s still in the early days, at least for mainstream adoption in the corporate world. That said, we do see a real opportunity here for leaders in the field to gain first-mover advantage. Our research suggests that at this point, when we look at CXOs, only around 20 percent are actively using or piloting any generative AI use cases. The majority are investigating it but not actively using it at this point. So there is a real opportunity here.

Martin Rosendahl: Now, the interesting question is: Is there a recipe for the reset and reimagine approach? I think what we have learned is that the reset really needs to be time-bound. We’re talking here about initiatives, like demand management requiring no investments. They are tough decisions, but they should be made fast. We’re talking about some organizational optimization. We’re talking about some external spend. We recommend doing all of these within a maximum of six months, sometimes even in three months. Reimagine, of course, is a longer-term effort. This is a 12-, 18-, or 24-month journey, with very tangible business benefits. So don’t compromise on what has worked in the past with business-driven use cases, and build for scale through the capabilities and technology foundation.

Matt Jochim: On the recipe, I think one of the other learnings is that it is really important for the CFO, the CHRO [chief human resources officer], and the CEO, as the triumvirate, to align and lead. As Martin described, these changes inevitably come with organizational redesign implications, questions about the grading of roles, questions about the appropriate structure and how to think about the matrix of the organization, and so on. So it is very important that those three roles that are leading the organization are aligned.

Another way to look at the generative AI question is to think about what’s the threat of not getting started. We had a conversation with a client recently, and we were thinking about exploring the potential application of generative AI for research and development and new-product innovation. The conversation was an interesting one. The head of research and development was at first a bit lukewarm to the idea of how much benefit there would be to his organization. But at some point in that conversation, it tipped into, “What if our competitors develop this capability?” And there was a bit of heavy silence. The reaction was that if they were able to generate products faster, put more ideas in front of the customer, that could actually be a real detriment to our business. So I think there’s a lens here, which isn’t just, what do I do with it, but what happens if my competitors figure this out? Is that something that would really put my business at risk?

Samdruk Thomas: Yes. And there are two different strategies, which Martin mentioned. One is a more defensive approach, which is investigating, how will AI affect my business and disrupt my ways of working? But the other one is a much more proactive approach where we think about, how can I identify sources of value to invest in?

Daphne Luchtenberg: And once you’ve got your strategy in place, how do you go about integrating digital into your G&A functions? We know that transformations are hard. What are some of the tips you can share here?

Matt Jochim: I think that one of the big steps is to demystify the whole notion of digital transformation. This isn’t about shiny tools and the latest and greatest technical advances per se. What it’s really about is: How am I going to run my business better? How am I going to do things at a lower cost? How am I going to do things faster? How am I going to do things at higher quality? This requires, as we’ve mentioned before, not just the technical piece, not just the systems and the data, but also new ways of working, new skills, different decision rights, different alignment around risk, and that sort of thing.

So I think the biggest breakthrough here is to think about digital as an enabler for changing, for reimagining the way that we’re going to manage SG&A, the way that we’re going to manage the business, but then also thinking about this as a business problem enabled by technology or a business opportunity, rather than thinking about this as something that the CIO needs to fix.

Samdruk Thomas: So when we look at what are the biggest success factors in making such a transformation successful, what our research shows is that actually one of the biggest enablers is having a clear strategic road map in place as to where are the pockets of value and how will we get there. Building on Matt’s point, many clients that we see are overindexing on the technology part, whereas it’s also about, how do we ensure we have the right operating model in place? How do we build up the internal capabilities, the change management approach, etcetera?

Daphne Luchtenberg: Building internal capabilities and the need for change seems like a great point to wrap things up. Many thanks to all of you for your time today. We’ll be continuing the conversation about SG&A cost management in another episode of McKinsey Talks Operations, looking at what we can learn about these functions from start-ups and disruptors. Does their approach add new opportunities for better cost management?

You’ve been listening to McKinsey Talks Operations with me, Daphne Luchtenberg. If you liked what you heard, make sure to subscribe and stay tuned.