Fast track to execution: how AI-First CFOs create value - ILOG SA
Hi, welcome, and thanks for joining our webinar today from IBM. We're speaking about our latest 2026 CFO study, Fast Track to Execution, How AI-First CFOs Are Creating Value. My name is Haynes Cooney. I'm a research director in IBM's Institute for Business Value, and I lead our global C-suite study series. I'm really pleased today to be joined by three terrific IBM colleagues for a discussion based on our study findings. This is Spencer Lin, our global research leader for CFO and finance in the IBV, and he's going to lead a discussion of the insight from this year's report. Josephine Schweilach is executive director for finance, data science, and technology for IBM Finance, And we'll be speaking to how IBM itself is becoming an AI-first organization, something we also like to talk about as Client Zero. And then Monica Pruthi is Vice President and Senior Partner and our Global Finance Transformation Lead for IBM Consulting. And she'll be sharing what we're hearing as we work with clients around the world to help them create and accelerate value. Hopefully together, our goal is really to bring these study findings to life for you to help you understand how IBM is applying them to ourselves and to think about the next steps you and your organization will be taking as we all move together towards an AI-first future. Before we dive into the report, I'd like to just provide a little bit of context. The IBM Institute for Business Value delivers trusted technology-based business insights by combining expertise from global leaders and practitioners, industry thinkers, and subject matter experts with global research and performance data. For more than 20 years, our thought leadership has provided timely and actionable recommendations to address the most pressing challenges and opportunities. And we're proud to say that the IBM Institute for Business Value is ranked the number one consulting thought leadership producer worldwide by Source Global research for both quality and impact. Since 2003, our C-suite study series has explored the priorities and challenges facing leaders at the very top of business and government. These reports reveal not only how these executives are leading change, but how leadership is changing. This year, we completed our role-based reports with a study we called the enterprise in 2030, and that really set the stage for a series of successive reports. It defined five characteristics of organizations in an AI-first future. Our CEO study then identifies five strategic plays CEOs are making to rewire their organizations to move with clarity and speed. The CFO study we're talking about today responds to these strategic plays and identifies the steps CFOs are taking to create value and execute the CEO agenda. Together, the research delivers a clear through line from the enterprise of the future to the steps CFOs are taking today to position their organizations for success. As I noted at the outset, the team we've assembled for you here today, with myself, Spencer, Josephine, and Monica, is really here to make the research even more practical and applicable by linking the study findings to the actions IBM Finance is taking and what we're hearing from clients with whom we're working. With that, I'm pleased to turn it over to Spencer Lynn to talk more about the findings of our 2026 CFO study. So, Spencer, over to you. Thank you, Haynes. So let me set a little bit of context for this year's study. So we surveyed more than 1,500 CFOs across 33 geographies and 26 industries and actually did additional supplemental data with interviews with a number of finance leaders worldwide. Now, what we saw across markets, industries, geographies, is that one message was very consistent across the CFOs, that the CMO mandate is expanding as execution becomes more consequential. this quote from Kawamura-san, who is the CFO of Kioxa Holdings in Japan, captures this shift around the CMO mandate. Although my business card says CFO, the reality is that the role is much closer to being an advisor to the CEO. So we're not saying that stewardship, financial stewardship changes. It still remains essential, but CFOs are now helping leadership, their C-suite brethren, connect choices, investment, and execution. So this is the central idea in this year's CFO study. Our research identified five execution disciplines that distinguish what we call AI-first CFOs. And together, they form what we called a CFO operating model for value creation. These are not five independent initiatives, and they are not a maturity model. They reinforce one another, connecting the overall strategic ambition of the enterprise investment decisions, and execution outcomes. What we saw is that organizations led by these AI-first CFOs reported revenue growth rates that were 23% higher than their peer organizations from 2022 through 2024. Now, we also saw that it was very consistent. We had an overlap in both the CEO study respondents and CFO respondents, actually nearly 1,200 CEOs were from the same organization that responded to our CFO study. And again, we saw very similar results of revenue growth outperformance for those CEOs that were part of AI-first organizations. Now, what we saw is that the difference, this difference in revenue growth is not really any single discipline, but how effectively organizations strengthen and connect all five of them. And the payoff that we saw actually extends beyond revenue growth. AI-first CFOs are more likely to report clearer enterprise direction and more consistent strategy execution, and they approve funding for new AI initiatives 15% faster than their peers. So what these outcomes have in common is all around execution. AI first CFOs create value, not because they invest more in AI, but because they connect enterprise advantage, governance, intelligence, capital allocation, and future options into a single operating model. So let's look briefly at each discipline, and then we'll transition to Josephine and Monica to talk about how these things are being applied within IBM Finance, as well as with our clients. So the first discipline is called Shape Enterprise Advantage. And this is really around helping make strategic choices executable. So finance aligns leadership, capabilities, resources, and measures of value behind the chosen priorities that are being made. And so it's really around helping the organization concentrate execution capacity on the opportunities that matter most. Now, one of the things that we saw is that this discipline becomes clearer when you look at what AI-first CFOs are doing. Rather than evaluating business cases after key decisions have been made, they're engaging earlier, they are helping leadership determine which opportunities deserve enterprise commitment, and they're figuring out how do you align capabilities and resources behind them. They're also more likely to help shape business model experiments. So enable new offerings and build strategic partnerships with their technology leadership. So the key difference for AI-first CFOs is that finance is now moving upstream from measuring value after the fact to helping create really the conditions for value before capital is even committed to opportunities. So that's what the research tells us. Haynes, let's now bring Josephine and Monica into the conversation and explore what this looks like in practice. Absolutely. Thank you, Spencer. So Monica, I'd like to start with you. And Spencer spoke to the need to engage CFOs early and to move upstream. When you're working with our client CFO, what are the clearest signs that finance is entering the conversation too late? And what differentiates organizations that are more successful at prioritizing enterprise-wide opportunities? Haynes, that's a great question. So for me, I think the clearest sign is when finance's first look at an initiative is the business case. Because by then, the strategic choices are already made, and finance can only approve, cut, delay. But you also see when the CFO is asked, well, what will this cost? But never should we be doing this. So what sets these successful organizations apart is when finance helps to define the agenda before the capital even moves. So just at the start, business partnering at that strategic level. The study shows that the AI-first CFOs, they help identify when which business models, value pools, ecosystem opportunities, and AI-enabled growth platforms warrant enterprise-level commitment rather than reviewing business cases after the fact. So in practice, that means, you know, looking at an enterprise-level value bet, sizing value across the functions, and having the discipline to say, maybe this one's not a good idea. Like, let's look at what are the best ideas, what's driving value across The Enterprise, and what's really important is to be able to then do that at speed and with agility to beable to pivot. Fantastic. Thanks, Monica. And Josephine, coming to you, Monica is talking about the evolution of the CFO's role here. How has IBM Finance evolved from measuring performance to actually helping shape where and how IBM creates value? Yeah, and let me align with everything Spencer and Monica said. I lead technology and transformation for IBM Finance internally. And I will be honest, historically, like CFOs, you know, at other companies, our role historically was to measure performance and report it, right? And we have moved to a place where we are now in the conversation before decisions are made. And the way we did that at IBM was very data-based. We used a tool, which happens to be an IBM product, called AppDio EBM. But what it does is it gives us full visibility into the costs and impact of AI throughout the life cycle of all of our workflows at IBM. So we did a large scale workflow transformation where we took all the activities needed to run the company and we mapped them into 10 key enterprise workflows. And then for each of those companies using this AppDio tool, we were able to ingest what does it cost to run that workflow at a pretty granular level. you know, going down, think of record to report is like the top level finance workflow. And then beneath that, we have additional granularity of a taxonomy of how the work gets done. We know the human cost, we know the technology cost, we know that it costs, including looking at things like token expenditure, which is probably near and dear to every CFO right now looking at the cost of tokens. But it's not just how much are we spending on it? It's what business outcome is it producing. So we look at ensuring consistent quality metrics remain. We look at the outcomes. So we're then taking this tool and mapping investment, adoption, productivity, quality, business outcomes. And what that allows us to do is it allows us to decide where to invest next. Because we can see for these workflows, where have we underinvested in AI? Where have we invested and it's a home run? Where Have we invested? And maybe we need to consider, you know, shutting down an AI tool that's not working. So we have 155 AI solutions in our workflows right now. And those are kind of the big enterprise solutions, not your personal productivity plays. And we continue to grow that. So finance owns this tool. Finance develops the data, and we now have the tools, the credibility, and the influence to figure out where should we be allocating capital. So it becomes very much a finance using data to get ahead of the conversation and help influence investment decisions. Fantastic. Thanks, Josephine. All right, Spencer, let's go back and take a look at execution discipline number two. So, the second discipline is what we call scaling AI with trust. And this is really the CFO's role around designing guardrails that enable scale. So as AI moves from generating content and insight toward recommendations or actually executing actions autonomously, governance really can no longer be treated as a policy or a compliance checkpoint. So the objective here is not to eliminate risk, but it's to make intelligent risk-taking repeatable at scale. So finance really needs to help define where AI can act, where human judgment remains essential, and who's accountable. So let's look at what AI-first CFOs are doing. And so what they're really doing is treating governance as an operating architecture rather than a compliance activity. And you can see here that they're establishing clear financial compliance and ethical thresholds. They're assessing second-order impacts of AI transformation, and they're creating shared accountability across finance, technology, and the business. And some are even allowing AI to play an active role in their investment and portfolio decisions, but obviously within defined financial and risk guardrails. So here it's governance becoming really a mechanism for confident action and not a reason to slow it. So the research tells us that governance has to scale with AI. Haynes, let's bring in, again, Josephine and Monica to explore, you know, how do organizations make governance an enabler? Absolutely. Thanks, Spencer. And yeah, this is really a compelling one. I think initially we heard from a lot of clients that they were nervous about using AI and some of their core processes because of these trust issues. But Josephine, how has IBM approached governance in a way that builds trust while still allowing AI adoption to move quickly and scale? So this is a good one. I like this because I think the answer now that I'm going to give is different than what I might have expected to give a few years ago. Because at IBM, we're 115 years old. We're a mature company. And you do get that way by being conservative, to be fair. But at IBM, we treat governance as an enabler, not a break. I would structure it when I talk about it with my team as kind of three considerations. And first is that governance has to be risk-based, right? All AI risk is not equal. A personal productivity use case, which is relatively straightforward, does not need the same level of rigor as, for example, my team has an AI-powered agentic journal automation tool, right, The level of rigor we need in that system is extremely high because it's material, right? It's important. It's critical that it be well-governed. So we provide nuance. And how we support that level of nuance is we have a centralized agent called Ask Integrity, where anyone at IBM who's creating an AI tool can use this to figure out, kind of triage, what level of risk does this have? And then they know to go to the CIO for this, to business controls, maybe to a team like mine, depending on the granularity. So let's first is figure out the appropriate level of risk. The second is make the safe path the easy path. And so I'm proud of this example because my team did have a leadership role in it. And this is how we use guardrails around AI coding for all of our finance professionals. So go back like maybe beginning of this year. I report to IBM's controller, which historically would be a pretty risk averse person. And he gave me the challenge of Josephine, I want every finance professional at IBM building their own AI, right? You know, I want them using our technology, I Want them using Our trusted governed data, and I want them innovating with solutions. And I kind of thought, Oh, my gosh, how am I going to do this, right, we have we have a tool at IBM called Bob, which is an IBM product, it's like a coding assistant, it's a software development lifecycle manager. And this coding assistant we wanted to make it the tool that all finance professionals could use. So we created something called finance mode, which combined guardrails, these risk-related concerns, but also accelerators. Nobody was going to use this mode if all it contained was guardrailes. So, we combined the two, right? We gave them dashboard accelerators, formatting, analytics, baked in along with guardraills. And so, that clear responsibility of the average finance professional to understand and govern the data while they build AI solutions helped us keep that trust in it. And then the third point I remind my team is that governance does not end at launch, right? You have to have a full end-to-end lifecycle where people understand they are responsible for their assets, they are responsible for confirming the accuracy of their assets. Responsible scaling requires ongoing ownership. So with those three kind of tenants, our objective is not zero risk. It is obviously to manage risk appropriately and aggressively, but we want to give people a clear, responsible path so they can move quickly and not have to figure out the governance themselves every time. Makes sense. Makes sense, Monica. What practical approaches have you seen our client organizations using to balance speed, accountability, and risk as AI scales? So I think the organizations and the clients that I kind of see this doing, doing this well, and they don't treat governance as, you know, a gate at the end. They design it, as Josephine was saying, as part of how work runs. So there are a couple of probably practical approaches that stand out. The first is around tiered autonomy. So organizations that are leading, they classify decisions by risk and materiality. So let's think of it as like letting AI act on its own for low risk work, like reconciliations or invoice matching. That's, I'm seeing the most there. Some organizations are, you know, using it to recommend for mid-tier decisions, and then leveraging the insights. I think there's a lot more underlying before we even get to the governance side around the data, which we'll hit there, probably in the next piece of this. But But then the third is around how you kind of leverage AI to help with some advisory, but there's much more on obviously the people side there where any fiduciary or capital exposure is high. So I think where I'm seeing this balance with governance is working the best with our clients, at least where they are today, is probably around that acting as low-risk work where AI can kind of act on its own. The second is around this making decisions rights explicit. So Josephine talked about this a little bit. It's like, who owns the outcome? When does the AI escalate and to whom? The ambiguity is really slowing things down. I don't think it's the risk itself. I think it is actually more the ambiguity around there without this accountability and ownership. And then as we think about kind of expanding autonomy based on evidence, so like a use case earns more independence as it shows accuracy and the audibility over time. Again, we're not quite there, but I think that's what we're trying to get to. So it's a little bit where like the study refers to this as engineering governance as an operating architecture, not a compliance layer. So as Josephine said, even with our controller, it's natural for finance leaders to think about how to run controls, thresholds, delegation of authority. So it's national for this to kind of sit with the CFO, and we're starting to apply that same thinking to our agents. Terrific, Monica. Thank you. And Spencer, Monica gave us a preview into what the next execution discipline is. Let's talk about number three. Yes, thank you, Haynes. Yes, the third discipline is around embedding intelligence into decisions. And this is really around the CFO as the architect of the enterprise decision system. So I think we can recognize that intelligence itself is becoming more and more abundant. Organizations can produce more data, more forecasts, more scenarios, and AI-generated recommendations than really leaders can absorb. So the real differentiator then becomes, can the enterprise connect trusted intelligence with the accountability, the accountable decision owners, and really act at speed? And I think this is why the core challenge, you know, really often is not about data availability or, you know, model capability. It's around decision design. So what exactly are AI-first CFOs doing differently to embed intelligence into the decisions? As I said, you Know, most organizations are really learning how to use AI, you But far fewer have really redesigned how decisions get made. And the AI-first CFOs, they stand out because, as Monica said, they're connecting trusted data, real-time integration, AI readiness, and workflow adoption into the decision system. So, the distinction isn't having better technology. It's about embedding that intelligence directly into planning, forecasting, allocation, and performance discussions where the decisions are actually happening. So ultimately, the goal is not more insight, it's better decisions. Haynes, let's now talk again with Josephine and Monica on what it takes to make that shift in practice. Yeah, thanks, Spencer. I think that's a really important one, right? And echoing your theme at the outset, these are not all discrete elements, but it's really the integration of a number of things. And so, Josephine, as we're thinking about embedding intelligence into decisions, IBM has invested significantly in integrated data, planning, forecasting, and AI. What has changed in finance as we've moved from producing reports to really embedding intelligent in everyday decision making? I think it's a great question. It's definitely tied to the first conversation we had, the first question, where we were asking about, you know, moving finance upstream, getting into making the decisions. I think what this question really gets to is how do we transform the actual decision-making process itself? Because you're now embedding AI into that process end-to-end. So one thing that I think is critical is that these reports we're creating, they're not the product, right? The decision is the product, the business is the product. And so historically, finance spent a lot of energy assembling information, reconciling it, producing a report, handing that report off for someone else to interpret. And so what we're asking now is the question, how do we bring the intelligence directly into the moment where the decision is being made? And I think one good example of this is around how we automated touchless forecasting in IBM Finance, because we have a pretty highly accurate data science model that we use for expense forecasting. And in our first iterations of it, we created reports that were consumable by the leaders, and that's great. But it's a very different experience than having the data science forecast embedded into where people are already doing the work. And so that's what we do now. They have a dashboard where they're looking at their sort of agentic variance analysis on the different workdays following quarter and month close. And we've embedded touchless forecasting into that experience. So now when you're assessing, hey, is this an outlier? Is it accurate? Should it be included? If it is an outliner that's accurate, we're pushing it straight into the forecast. And then that forecast can be consumed right there. So it really changes the role of the finance professional because the technology is doing the assembly, the reconciliation, the first level analysis, writing draft reports. And then the analysts are right there applying their judgment, challenging the assumptions, partnering with a business on what to do next. I will add as a part of that, it's not just about embedding the AI output, but you also have to embed the trust and transparency and explainability. So in that touchless forecasting example, we brought in transparency, simplified models, the ability for the analyst to challenge the model and understand the underlying data science. And that got us to a place where embedding intelligence is not just about better reporting. It's about making the information available at the time decisions are being made, and then redesigning the workflow so that that's most effective. Very nice. Thanks, Josephine. Monica, building on what Josephine is saying, I think when we do survey research, when we go out and talk to clients, oftentimes we hear them saying they have a data problem. In your experience, how often is this, you know, is the real underlying issue a workflow or an operating model or a decision design challenge instead? I mean, I love this question because I haven't met a client yet who hasn't said they have a data problem. I know data is at the foundation of this. So, you know, most often than not, when a client says we have a data problem, it usually means like data exists somewhere, but no one's really designed where it enters a decision, who acts on it and how fast. And this is where you get to the, is it a data problem or is it workflow operating model decision design challenge, as you said. So a good test is maybe saying to a client, if your team had perfect data tomorrow, would the decision change? And would it change faster? And if the answer is no, then maybe it's a workflow or an operating model issue. Some of the common culprits that we have, even we had with ourselves, and I see this all the time with clients, it's handoffs between finance and the business. It's approval cadences. It's accountability that's spread so thin that nobody owns the decision. The studies saw that 6% of the finance organizations operating at a level where AI consistently shapes decisions at scales. That's only 6%. The leaders are redesigning the process so that the insight is produced and it's acted on continuously instead of layering AI on top of legacy workflows. Data matters. Data is the foundation. But if the question is, I have all the data at my fingertips, then it's the decision design that's usually the unlock for how to actually get a lot of the real value. I love that. I Love that. So Spencer, we're going to come back to you. We're hearing about making decisions, better systems for decision-making. But then what happens? How do you actually, how do organizations actually take action on the decision once it's made? So let's talk about execution discipline number four. So the fourth discipline, which is a little bit of a shift, is around enabling dynamic capital allocation. You know, clearly a very important area for CFOs of being the orchestrator of capital motion. And I think both Josephine and Monica talked about AI being used to identify opportunities and risks, and obviously it can do it continuously. But many organizations are still funding AI through periodic planning and a fixed approval process. So, you know, that result is a growing disconnect. You know, you have insight moving in real time, but capital often does not. So the objective here is to create, you know, a capital system that can sense, decide, fund, and reinvest and stop as, you know, evidence unfolds. So, you know, clearly AI has created a new challenge, right? Insight can now move faster than capital. But what AI first CFOs are doing is they're trying to close that gap. You know, they connect, you Know, as Josephine mentioned, value tracking, real-time visibility, flexible funding, and really reinvestment mechanisms so then resources can move, as I mentioned, as evidence changes. And I think more importantly for AI-first CFOs, they're simply not just using or harvesting AI-driven productivity as savings. they use it to create investment capacity. You know, and our own CFO talks about that creating a flywheel, right? Productivity generates capacity. Capacity funds investment. Investment enables growth, and then growth creates really the next wave of opportunity. So, you know, our research really suggests that execution speed really increasingly depends on how quickly capital can move. So, Haynes, let's now talk about what it takes to build that capability and practice. Absolutely, Spencer. So, Josephine, I'll come to you first again. Spencer mentioned Jim Kavanaugh, our CFO, and he talks about a flywheel. Spencer described it, productivity is creating capacity for investment, investment funds the next wave of growth. How are we thinking in IBM and IBM Finance about reinvesting productivity gains rather than simply harvesting them as savings? Yeah, absolutely. I mean, at its heart, productivity creates the capacity that we need to invest. You see that at the enterprise level. We've gone public saying over the past few years, we have created what this year will be a $5.5 billion productivity run rate. Those are hard quantifiable dollars. We achieved 4.5 million last year, and we are on track this year for 5.5. And that capacity has allowed us to invest back in the business. We prioritized what matters to our clients. So we increased R&D by 50%, including investing in cloud, AI, quantum, semiconductors. We expanded our technology sales capabilities to better serve our clients. And then we invested in acquisitions, $20 billion in 38 acquisitions focused on cloud and AI. And we call it a flywheel, right? Like many tech companies, we love talking about a flywheel. But we create that productivity. Some of it gets harvested as savings, and then some of it gets reinvested. So that's at the enterprise level. Now, I said at what we call the domain level. I said the finance level. Within our own finance business, we have to make that same decision. When we're driving productivity by building these new technologies that my team is releasing, we want to use some of it to reinvest in what the finance analysts are doing. So we have be very deliberate. What work are they going to stop doing? What higher value work is going to get expanded? What skills do we need and where do we want that capacity to go? And as you think about capital allocation, even across, for example, the CFO team, I think it's really important to think about change management. So I'm going to go back to the touchless forecasting example I gave, because I want to be honest, when we rolled out that highly successful change management, highly successful touchless broadcasting tool, the first six months, everyone hated it, right? And no one wanted to use it. And the reason for that is because we hadn't leaned enough into change management, into A, making sure that we provided all the trust and transparency tools that we enhanced. But B, to the heart of this question, we didn't tell them what to do with the time of read up. And so what happened at first is people were using touchless forecasting, but then they were running their legacy process in parallel because they still had that time because we hadn't been clear about the higher value work they should focus on. So that was a big change management lesson for us was ensuring that all leadership from the first-line manager to the CFO communicates with these new capabilities, you're going to now focus on scenario planning. You're going focus on headcount strategy, right? Whatever it is for that individual person. And I think that's key. And that's what makes it a flywheel rather than a cost program is create the capacity, reinvest it, and then harvest it to make the business stronger. Yeah, it makes sense. And I really like that point around making sure you've got clear guidance on the back end so that the team is focusing on your highest priority things. So, Monica, coming to you, what prevents most organizations from moving capital as quickly as new opportunities emerge? And what capabilities are distinguishing those organizations that are doing that successfully? I think we just heard right from Josephine on this. You know, you've got to be able to take that productivity and actually it's creating the capacity that needs to be reinvested. to fuel the growth. I mean, that, that is definitely where, you know, the, the real dollars are. I think the barriers are probably structural, maybe not analytical, like, you know, like the annual budget becomes an entitlement. It's kind of moving money from, you know, one to another. So it's, it becomes this negotiation. The approval processes are kind of built for predictability. So, you know, then we look at like, well, there's a mid-year shift and then there's several committees and it's like it just it's this lag that's taking a long time a lot of the organizations can't see the value in real time and they're also still not looking at it from an enterprise-wide value level it's still stuck in these kind of you know siloed programs so you don't necessarily see the evidence to justify and a moving capital around um you know where you can so the the organizations that move faster they share probably a few capabilities like they're as you heard from josephine right we we talked about the productivity that's creating the the capacity that's fueling the growth like that's definitely one it's it's also funding outcomes um instead of projects so this goes back to the enterprise like what is the outcome that you're trying to fund um and then some organizations they you know they might even hold capital in reserve you know, without kind of having to redeploy or reopening a whole plan that this way you can kind of redeploy quickly. I have seen that within some organizations there as well. You know, it's really about how fast can the resources follow the evidence that they're seeing. So, you know where this study said 84% of AI first CFOs, you know, they have this shared dashboard and it's something that we have too. And it's certainly something we've built for clients and we've, we've seen clients have that, you Know, everyone's working from the same view. So you're able to see what's working and that that's an advantage there. And so therefore the resources can follow the evidence pretty quickly because once you have a view like that, and then you apply the AI and the insights on top of it, you can actually redeploy pretty, pretty quickly across that capital. Very nice. Thank you, Monica. So, Spencer, let's come back to you for the fifth and final element here. So let's talk about the future. So the fifth discipline is what we call building future options. Now, in the fourth discipline, you know, dynamic capital allocation. That's really around moving resources toward what's creating value today. In this discipline, it's around preparing the enterprise for what may create value tomorrow. And because obviously the future can't be predicted precisely, CFOs really need mechanisms for funding experimentation, learning quickly, and deciding, you know, what to stop, scale, or redirect. So, clearly, in volatile environments, you have this experimentation becomes a financial capability. And what are AI-first CFOs doing differently? You know, they're really around treating this notion of optionality as a managed capability rather than, you know, a collection of isolated experiments. So what exactly are they doing? They're using future scenarios, stage funding, stop scale mechanisms, and shared accountability really to create a repeatable system for the learning, the scaling, and the adaptation. And I think the distinction between this discipline and the previous one is really important, right? Because the dynamic capital allocation is really around optimizing today's portfolio, whereas building future options is around expanding the choices that are available tomorrow. So the goal obviously is not to predict the future better. It's around building the flexibility to respond as different futures emerge, whether it be quantum computing, advanced forms of AI, et cetera. So it's less about predicting what happens next and more on really having the ability to respond. So, Haynes, let's now talk about how organizations can build that flexibility while still delivering the results today through their dynamic capital allocation. Yeah, it's great. Thank you, Spencer. So, Josephine, Spencer talked about systems that enable flexibility on sort of a repeating basis. With so many AI ideas competing for funding, how does IBM Finance decide which projects to double down on and which to shut down? Yeah, I think it's a great question. I think in some ways this is the easiest question because all we do is we treat our AI portfolio like a managed portfolio. And we scale the investment to demonstrate value. We reuse them and we direct resources away from work that's not showing a path to value. It's obviously not that easy. So let me be a little more tactical on what we do. When we start experiments, we have clearly defined success criteria and a clearly defined timeline. So say, for example, right now we have in production our agentic variance analysis for close. When we started that project, we had 30, 60, 90, 120-day checkpoints that said, if at this point we cannot X, then we are going to disinvest. Because I think we're all, as you know, all companies are facing the reality that we all have, even at IBM, we have a small pool of AI engineers who can work on these projects for me, and we have to use their time effectively. And so having clearly defined criteria on when to stop is really important. It's a difficult premise, but not every good idea is going to scale. There's a lot of variables. We use a pretty complex offering management methodology to prioritize my funnel. I mean, I have 100 requested projects that I'm deciding between. One thing I would share with my peers at other companies is that one of the variables that matter is what I call client readiness. And that means that not every team within our organization is equally ready to adopt change, regardless of change methodologies. And so there's justification for saying this organization needs to change, they're ripe for it, and their senior leaders have committed when we built this technology, they are going to use it. And by the way, there's some first-line analysts or finance employees who are excited about helping us build it. That's when we know we're going to get to value. So I think really it's at the heart of very simply treat it like a managed portfolio. I mentioned earlier, we use IBM's value analysis tool to look at all the aspects of our AI projects. Are they accurate? How much do they cost? How much are they impacting different workflows? So we have a hard metrics we use to manage the portfolio. But there's also just gut, right? We all know when an organization is ready for change, then it can be worth making that little extra investment. Well, I like that. I like the combination of on the one hand, you know, this is what you know how to do already. This is what we do in finance, but also recognize some new elements like this AI readiness that you talked about and how that may impact your methodology. Monica, many of the client organizations we speak to have dozens of pilots that they're investing in that they are attempting to implement. What distinguishes a managed exploration portfolio from just a collection of disconnected experiments? Yeah. So, Haynes, like, no organization is short on AI ambition. I think the real question is, is how are some able to turn this into enterprise-wide business outcomes and see big value, while others are really stuck in these pilots and these fragmented initiatives, like from this collection of disconnected experiments. The CFOs that are pulling ahead, they treat AI as an enterprise-wide shift. They're tying investments to end-to-end outcomes, and they're redirecting the capital as value emerges. So these organizations are growing 23% faster than their peers, as we've seen in our study. I think, as Josephine, and my answer is very similar just beans it's about the stop start stop start culture here you know a managed portfolio has many hallmarks like every experiment is tied to an enterprise hypothesis it's not a team's curiosity this success and stop criteria are set up front with a time box that is what where most organizations are really struggling is that's it's actually the stop criteria um you You know, there should be a regular forum that actually makes the stop, scale, pivot calls and recycles the funding. It's how we're guiding a lot of our clients to actually set up that governance. And as I said, the stopping is the underrated part, like when to say this isn't working. So organizations that end pilots cleanly, they free up capital, they fre up talent and then redirect to focus on higher value added things. So, you know, finance is really, they're well-placed to run this because it's essentially portfolio management applied to learning and understanding kind of where to redirect the funds and the talent. But, again, that stops our culture. Fantastic. Thank you, Monica. And so we'll move here, and we'll just quickly tie a bow around the sort of a couple questions that CFOs can now take to sort of address the question, where can leaders start, right? So drawing on the five execution disciplines we've discussed today and what we've been hearing from Josephine about how IBM is addressing this, from Monica about how many of our clients are addressing this and the kind of guidance that we're providing, Here are five questions CFOs should be asking their teams. So, number one, are leadership capabilities and resources actually aligned behind the opportunities most likely to create long-term value for our organization? Two, do we have the AI guardrails in place that enable intelligent risk-taking, or are they really just sort of slowing adoption, and how do we find that right balance? Number three, what high-value decisions remain constrained by fragmented data or manual reconciliation, unclear ownership, this failure to connect those dots that we've been discussing today? How and how quickly can productivity gains and investment capacity be redirected toward growth? What's standing in the way of pursuing our opportunities? and fifth and finally speaking to this last point does our organization have a managed exploration portfolio or are we operating a collection of disconnected pilots to monica's point what should we stop and where do we need to focus on going forward where the answers to these questions reveal barriers to execution and growth and what actions need to be taken by whom and when and in sequence to maximize value? So I'll pause there. I think we do have at least one question in the chat, and then we will certainly invite others as well before we wrap things up. So I see a question. I'll ask this, and maybe, Monica, I'll direct this to you first. So for a government client in the public sector, any suggestions to define value in a governmental situation when the goal isn't to increase revenue, but rather to improve how we serve constituents. I also like how you said maybe I answered my own question there. The rest of that question. I mean, I don't, it's like you said, it is not about increasing value, but it is how to improve. So maybe it's probably more focused on things like governance and data and prioritization, but it's not necessarily on, you know, driving that, how do I fuel the productivity to drive the additional growth? It's more about how do you ensure the right guardrails are probably there in place. And that'll then kind of, I think that'll then help your constituents as well. Because if you're able to get to data to drive insights faster, and then you're able to get to things that your constituents would need. Haynes, I think also, too, that if you start looking at specific dimensions, maybe it's around service quality? You know, how long does it take for a constituent to receive, you know, a particular service or benefit? You Know, what type of experience that are they having? Program outcomes? You know what does it cost to serve? That you know those might be some of the metrics or value type dimensions, you Know, that you would consider, right? Because I mean, if you use an example of i don't know ai improving the time um to process a benefit application right and if it goes from let's say 10 days to two that's creating value but it's not really generating any new revenue um and then you know second example would be you know automation helping you know your case workers from doing, you know, administrative tasks so that they can serve more people, then, you Know, that's also about creating value, but it's not around revoting. Yeah, I think that's a, those are great responses. And I would also point back, Josephine referenced something about having a real clear plan of what you're doing with that capacity that you've freed up. I think in a mission-driven organization, I think one of the benefits and having also spent some time in the public sector myself, folks are usually quite clear on what the objectives of the organization are in terms of serving constituents and providing that service. And as Spencer was just saying, being able to reallocate time from administrative tasks to the next thing, which is aligned with that mission is often an opportunity that's there. Any other questions from our guests today? Okay. With that, certainly don't hesitate to reach out to anybody on the call today. We're always happy to hear from you and would look forward to working with you to address the challenges that are in front of your organization. We'd recommend that, you know, continuing this conversation, engage with the study. We've got a QR code here as well as a link. Think about how your organization stacks up on each of the five disciplines that we've walked through today. and reach out, engage us, take a look at the finance consulting services page as well, and look at different kinds of examples and case studies that are featured there. And with that, I would ask each of our presenters today, if they have any final closing comments you'd like to share, just to focus on anything our guests should walk away with. So Spencer, I'll start with you. Yeah, again, I think the key thing to remember is that, you know, this is a CFO operating model, and the five execution disciplines are really interconnected. Thanks, Spencer. Josephine, over to you. Focus on quantifiable value. If you want to get to scale, you want to drive success, you wanna drive belief in your organization, it's really helpful to have a data fact base that you can prove success on. Fantastic. Thanks. And Monica, last word to you? I think I would focus on enterprise-wide, too. So think beyond just the finance organization and also forget about the people aspect of this. Perfect. Sounds great. Well, thank you all. Thank you, Monica, Josephine and Spencer for your insights. And thanks to all of our guests for joining. With that, we look forward to speaking with you next time. Take care all. Have a good day.
- Company
- ILOG SA
- Keywords
- ILOG SA