Building the path to agentic AI adoption in FP&A

Good morning, good afternoon, and good evening, everyone. Wherever you are in the world, welcome to our webinar, Building the Path to Agentec AI Adoption in FP&A. My name is Prash Charghi. I'm the Global Director of Planning and Analysis of Product Marketing at OneStream, and I'm joined by some amazing colleagues from Accenture that I'll introduce in just a bit.

But before I get to that, the purpose of today's webinar is really to help you as a finance and FP&A audience understand the path that you need to get to adopt an Agentec AI framework within your FP&A space.

We're going to talk to these experts from Accenture today about the readiness path, where you are right now, and where you need to be so you can have a full and pleasant and amazing Agentec AI experience going forward.

As part of our webinar today, we're going to go through a quick Agentec AI overview from my part, and then we'll have our experts that I'll call on in just a bit to have an AI-level discussion in terms of the building blocks, as I mentioned, the path to get there, some of the use cases that we'll review, and then finally we'll end with a conclusion and some Q&A from you, the audience.

So with that in mind, I'm going to ask my colleagues from Accenture to introduce themselves. So Jose, I'll start with you. Yes. Hey, Praz. Thanks for having us here. So my name is Jose Melendez with Accenture.

I've been in this space for about 17 years. As this has evolved, we've seen a lot of changes in the finance space. I've helped firms with transformations across both R2R, FDNA, supply chain integrated business planning, and so forth.

So looking forward to the conversation today as we look into the new horizon of Agentec AI. Fantastic. And Dhruv? Good morning, good afternoon, good evening, everyone. Dhruv Jain, also with Accenture, and excited to be part of this webinar, Praz.

Thanks for having Jose and me join this call, this webinar. I've been in the industry for over 30 years, and that will give you an idea of my age, which I don't like to talk about very often. But I've spent really the last eight years, eight to 10 years, working with finance teams on everything from data and analytics, and now AI, Gen AI, and Agentec AI.

I've worked with clients on how do we make Agentec real in finance, both in the finance space and in the adjacent space. And also, I've had the opportunity to work with our own, Accenture's own finance team.

So super excited to be here and have this conversation and learn from the questions you'll ask. And I'm sure the questions Praz is going to ask us. Well, thank you very much, Dhruv, and thank you very much, Jose.

It's a pleasure to have you guys here again, and let's get started. So right now, before we get started on our discussion, I'm just going to spend some time going through an overview of Agentec AI. So where we are is basically, there's a lot of excitement around AI right now.

And, you know, we have to think about why now. Now, if you've been in the finance space long enough, you've probably seen technology waves across the board, whether it's ERP, dating ourselves, you know, cloud planning or moving to the cloud, RPA.

Each one promised a sort of transformation, and it did deliver, but partially. So there is something that makes this moment different. And the honest answer is, it's not just one thing. It's really several things converging at once.

Now, the first thing is that, well, the infrastructure caught up. You know, whether you call it cloud, big data, advancements in chips that we hear about. It's really helped create a foundation that makes AI viable at scale.

The models also got dramatically better with LLMs that we hear about, and they can work with the kind of unstructured or narrative-heavy content that, you know, finance individuals that we work with all the time, they produce every day.

The also, another factor is that the economics have shifted. It's really no longer an experiment. Rather, it's an investment that has a quantifiable return as well. And most important, you know, what was once a competitive differentiator 12 months ago is really becoming table stakes.

Organizations that haven't moved are holding steady, but they're also falling behind. So this brings us to agentic AI. And it's worth pausing on because it's really a genuine step change. It's not a better version of what came before.

You know, the AI most finance teams experimented with in the past has reacted. You ask something, it responds. At agentic AI, it can be given a goal, executed autonomously across multiple steps to achieve it.

In FP&A terms, you know, imagine your monthly variance analysis, pulling data, normalizing it, identifying outliers, drafting commentary. That entire workflow can be executed by an agent, not just one single step of it.

That's the shift from AI as a tool that we use to a colleague that works. So it brings us to the urgency. Adoption is accelerating fast, faster than any technology we've seen. 58% of finance functions are using AI in 2024, up 21% in a single year.

The question isn't whether agentic AI is coming to FP&A. Rather, it's whether you're driving it or reacting to it. And that's what we're here to unpack today. So with that in mind, let's get over to move over to our first topic, the agentic moment.

Why FP&A can't afford to wait. So one of the things we know is that from our research with FP&A trends, only 9% of finance teams have fully automated core FP&A processes. So with that in mind, I ask Jose a question that might parlay off this data.

What's the most important distinction that you draw between what gen AI does, which we've heard a lot about, and what agentic AI does in plain language that a CFO would understand? Yeah, thanks, Pryos.

And you kind of touched on it in the previous element. Like the technology has been evolving, right? So terms like RPA, robotic process automation, cognitive RPA, machine learning, all of those things have been the evolution of where we are today.

And they've all been building upon each other. So where RPA was mostly scheduling and automating a manual task, machine learning brought automation and human judgment and was leveraging ML and DL models.

Agentic AI, like you mentioned, is autonomous, right? But it still leverages a lot of these core technologies that have been placed in the past. But the real shift is we're moving from descriptive analytics, which was kind of in the earlier level, to predictive analytics, to prescriptive analytics, to the level now that we're talking about, which is more of an intelligent enterprise, where, like you said, these autonomous agents are able to do end-to-end tasks.

They can be put in the background and be running to support a process or a function. Thanks so much, Jose. Okay. So based on what you just described and obviously the case for moving and really what gender AI to agentic AI is, where are you seeing the most urgency right now for finance leaders? Is it competitive pressures, efficiency pressures, or something else completely different that's really driving this need for change and this disruption? Yeah, so this is a great question because this is exactly what a lot of our clients, CFOs, and finance partners are dealing with, right? They're really looking at an evolving environment with four key disruptors.

Technological advancements, like you described, right? The technology is moving faster than most teams can adopt it. New competitors coming into various industries. The increasing cost of capital and shifting consumer preferences.

So all of these are the things that are top of mind to the finance organization, which is then asking themselves, how do I leverage these new tools and new processes to stay ahead of the curve? But the reality is disruption not only affects finance, but it affects the entire organization and the role finance plays within it as a partner across the org.

So no matter how you analyze the findings, these four core disruptors are common to all business and finance functions. But the difference lies in how they manifest themselves. For example, a consumer goods industry company, technology has a large impact on how the enterprise operates, for instance, their supply chain, which could impact integrated business planning with supply chain and finance.

In the services industry, you could see an example where technology impacts the type of experiences and engagements offered to the customers. So finance might be more affected by the relationship between sales and marketing planning and the role the finance plays in that.

So it's definitely top of mind. It's moving quickly. And we're all here to kind of help our clients think through how to adjust and adopt. Great. Thanks so much, Jose. So I'm going to take this stat that we have and flip it around a little bit.

I mean, you know, we see 9%, but the inverse of that is that 91% of finance teams still have significant manual work. What's keeping them there? Is it the technology or something deeper? Yeah. And this is the very interesting one is throughout a lot of the technological advances that we've heard, you're going to find that some of the similar barriers still exist.

So the top five barriers that we see are really lack of quality data, lack of technical infrastructure or legacy technology that these tools cannot be built upon, lack of skills, the talent that's within the organization, the speed at which the environment is changing, and sometimes insufficient availability of capital.

But the overall element also comes with these technologies become more advanced, there's also lack of trust in how this is all coming together. So all of these factors are creating barriers, but there are also opportunities that the org can address individually to ultimately break through.

Fantastic. Thanks so much, Jose. So with that in mind, we're going to move over to our next topic, the readiness reality, and really find out now where finance teams actually stand. In order to do that, let's look at our stat over here.

The number one, when we did our research with FP&A trends, what we found is that data quality is the number one barrier in AI. And we've heard this over and over again, data has been a challenge for not just FP&A, but finance overall for generations.

But it's at a point now where it has to be addressed. So with that in mind, let me propose my first question to you, Jose. So when you go into a finance organization and do an AI readiness assessment, what's the first thing you look for? And what do you usually find? Yeah, surprise.

I'll kind of answer this maybe fairly quickly around how we approach an assessment, right? Because that's ultimately what we're trying to do. We're trying to understand the current state, define a future state, and ultimately identify the actionable element.

So we use a three-pronged approach of explore, experiment, and execute. And ultimately, during this initial phase, we're looking for that readiness criteria across five key levers, data, technology, people, process, and ultimately identifying the value.

We'll identify use cases and define a North Star vision. We also want to focus on identifying the gaps and the barriers to scaling across the finance or the enterprise, whether that be siloed functions, poor quality data, et cetera.

So like I mentioned with the previous barriers, we usually come out of there with findings around the data, the process, and the technology, which are the common barriers that we find. Awesome. Well, you mentioned data quality.

I started with data quality. We just keep coming back to data quality as a top barrier. Is that a technology problem, a governance problem, or a people problem, or all three? I'd say it's probably a combination of all three.

And the reason why is data quality is typically referred to as the usability or the accuracy of the data as an input into a process, right? So whether it's finance teams doing things individually or as part of an automated process or as part of an agentic solution that might come in the future, those inputs have to have, quote, unquote, quality.

But the problem is that data originates across the organization. And typically, finance doesn't own a lot of the data that they use. And so they rely on other people, other processes, and other technology to feed that information to them.

And this is where the complexity lies. As a finance organization looks to transform their process or introduce these tools, they have to partner with these broader parts of the organization. And they have to take a primary role in defining the expectations of that data as inputs into the processes they need.

So it's definitely a complex challenge that finance teams have. But it's also the accountability of the enterprise to ensure that that quality of data is there because finance is not the only consumer of that data.

Makes sense. And one thing there, Praz, is, you know, even within finance, sometimes we see that the data requirements at a corporate level are very different from the data requirements at a business unit level.

Right. So working with a global pharma company right now and the challenge they're dealing within finance is how do we reconcile these two? Right. Because what we get at the corporate aggregate level is much too high level of rollout, an aggregated sum for them to run the business operation at the business unit doesn't have the transaction.

So it's also bridging this gap across corporate and the business units, which is super critical. Great. Thanks so much, Dhruv. And that leads me to you. So, you know, we talk about data quality, but, you know, where you are with data quality might often reside with where you are with regards to data maturity.

And there's often a big gap between what, you know, leaders across an organization think their data maturity is and what it actually is. So, Dhruv, how do you close that perception without effectively derailing the momentum? Yeah.

And I call this the classic duck in the water, you know, the duck in the water syndrome, right, is when you look at the duck gliding across the pond, it's just effortlessly moving from point A to point B from the shore to the middle and back.

But what you don't see is the furious paddling under the water for the duck to stay afloat, right? And I think of data in a similar fashion, right, is when leaders see it, they see, yeah, I get my reports on a monthly basis, right? I ask for it, it comes to me.

But it's the teams which have to actually pull the reports into the analysis. That's the ones who are spending a lot of cycles getting it. And where it's, you know, what my suggestion is, pick an area.

So, Jose talked about identifying value, right? So, once you've identified value, which has to be unlocked, start there, right? Start something which has got the highest value because that's also got the highest visibility and probably the highest air cover support from your leaders.

So, start there, identify what are the issues there, what are the problems there, and, you know, use that as a way to build the momentum to start not just cleaning up your data, but also reconciling and mapping your data.

So, going back to that example I was giving up with the Global Pharma Company is they wanted to get their data AI ready, right? And the value was they wanted to be able to use it for workflow orchestration to reduce the time which was spent.

And so, they got alignment with their CFO and their group CFO and their IT teams that we've got to go and fix this issue between corporate granularity and view granularity, right? And let's get the two to work together.

So, as an example, and instead of starting with all the data elements which sit in an ERP, they said, let's start with a given set of data which we know is controlled and managed. And then work our way through it.

So, start small, get some quick wins, and build momentum. Fantastic. Well, thank you so much, Dhruv. And with that, let's move over to agentic use cases that actually matter in FP&A and effectively matter right now.

So, what we know is that finance teams spend about 60% to 70% of times on time on data gathering and not on analysis. And, you know, it's funny, if we think about the profession, it's FP&A, financial planning and analysis, but how much time is actually spent on the analysis is very limited versus the amount of time that's spent on the planning overall.

So, with that being said, let's move over to the first question to you, Dhruv. If you had to pick one FP&A use case where agentic AI is delivering real measurable value right now, not in theory, but not on a roadmap, what would it be? Actually, there are three use cases that come to mind when I think about where agentic AI is delivering value.

Two of them are in FP&A and one of them is in accounts payable within the help desk space, right? And this is an example, and I'll go to the FP&A examples. But within the AP help desk, we are seeing by using agentic AI, we are able to triage vendor requests, vendor queries, and reduce the turnaround time from almost 30 minutes down to two minutes, right? Which is a phenomenal time saving, and also from a user experience perspective, right, is we can respond to the vendors in a more timely manner.

So, that's one example where we are seeing tremendous savings in time. Within FP&A, there are two places where we are seeing a lot of benefits. One is, I think, Praz, you mentioned earlier, right, is the FP&A teams are responsible for a lot of narrative-heavy content, right? Especially to support month-end reports, board reviews, things like that.

Where we are seeing AI, agentic AI, generative AI being very helpful is in writing these draft commentaries, right? And getting them ready for review by the FP&A teams. And it gives the FP&A teams a starting point to then go and have a conversation with the business partner.

So, you know, they're not spending the time doing the writing, but they're spending the time doing the analysis and the conversations to reach the discussions. So, that is where we are. That's one of the FP&A areas that we're seeing a great adoption of agentic.

The other place where we are seeing a lot of value being unlocked is in embedding a conversational capability along with reporting. Because today what happens is FP&A teams, when they publish, when they, you know, say, these are my month-end reports, these are the dashboards, this is the variance analysis between forecasts and actuals.

They provide that input in a PowerPoint deck or a set of slides. And then there's always going to be three or four follow-up questions. And this is where we are seeing agents come in and help is by embedding agents along with these dashboards.

People are now able to look at the number on the dashboard, ask questions, and, you know, do some self-discovery. So, they've already gone down two, three, four follow-up questions before they pick up the phone and call the analyst to ask them, you know, can you provide me more content? Or when they're asking for the information, it's more nuanced, it's more targeted, it's something which requires deeper analysis.

So, again, this is where we are seeing the agents take away some of the drudgery or the more mundane tasks and freeing up tasks for the FP&A teams to do the more critical, the deeper thinking, analysis, business partnering kind of activity.

Fantastic. So, I want to go back for a second to what you mentioned before, variance analysis. And that's something that stuck out for me from my FP&A days. And I still shudder at the thought of volume variances, rate variances, mixed variances, explaining the variances.

And we know it's one of the most time-consuming and repetitive tasks within FP&A. You know, walk us through what that actually looks like when an agent is doing the heavy lifting. Yeah. So, I'll bring that to life with a story from some work we did with a global retailer, right? And the global retailer operated in multiple business units globally.

And they sold their products. I think they had 10 categories on which they sold their products, right? So, they had over 50 business units. Now, the FP&A team, at the end of every month, would have to look at the forecast, would have to look at the actuals and do the analysis and prepare summaries, which would be reviewed by the BU leaders, right? So, the 50 BU leaders had to review it.

All the category leaders, right? All the regional leaders, both across the BU's and across the categories and then at the global level. So, it was taking the same information and summarizing and synthesizing it in five or six different ways.

And all of this had to be done in those eight to 10 days that this company was taking to close their books. So, this is where we were building out an agentic system where there were agents which were looking at the variances.

And depending on whether it was a BU, a regional, a category, or a global level, right? It was taking those variances and it was summarizing them and building out the commentary. Because, again, depending on which level you are at, depending on the size of your BU, what may be relevant for your BU or your category may not be relevant at a global level.

So, how using the agents to kind of support the manual effort, which our FB&A teams are today doing, right? And so, in a way, building out the hierarchy of agents to mirror the organizational hierarchy for reporting and to get the draft variances completed more quickly.

Excellent. Thank you so much. So, with that in mind, let's move on to our next topic. The human in the lead imperative. Governance, trust, and accountability. So, we know that 53% of workers, especially in finance, don't know who's accountable when AI makes an error.

And I think that's something that we all have to think about, especially as we adopt, you know, go towards more and more of, you know, from a gen AI perspective, especially to an agentic perspective where more and more tasks are being done, especially in the context that you'd recommended, Daru, before as well.

So, with that in mind, you know, what I'm thinking about, Daru, is human in the loop gets used a lot. But what does it mean in practice for an FP&A team deploying an AI agent? You know, where does the human stay and where do they let go? Yeah, so I'll paraphrase our CEO, where we refer to now as human in the lead, right, as the human plus agent, right? Because agents also, I look at them or we look at them as a workforce.

They're just digital. So, but where do humans take the lead? And in a very simplistic way, the way I think about it is humans have to take the lead for any work for which they're responsible and accountable for, right? If there's anything which they're accountable for, they've got to take the lead and drive that, right? So, with an FP&A example that I was giving, right, the agents can write the narrative, right? Or in another case, they can look at internal and external signals and, you know, they can develop a scenario, right? They can draft the narrative.

But which scenario to accept, which draft to accept, right, that the FP&A team is accountable for, which is why they have to look at that, they have to review that, and they have to say, yes, this is the scenario, this is the narrative which we want to push out, right? Because once they approve that, then it becomes a matter of public record, right? For a publicly traded company, the company becomes a part of your investor call.

For your scenario, whichever you accept flows into your planning, budgeting, PBF software. It informs your forecast, which you may eventually report out of the street. So, that's why if anything which you're accountable for is where humans need to take the lead, the other work can be designated to agents.

Perfect. Perfect. So, the other question I have to ask you, and I think it's one of the most important questions today. There's obviously a trust problem on two sides with leadership trusting the AI output and then the finance team trusting that they won't be blamed if or when something goes wrong.

How do you advise FP&A and finance and leadership overall to address, you know, both? This is a very interesting question, and this is something that we get asked a lot, right? Because especially in finance, anything that we do needs to be traceable, auditable, and explainable, right? Unlike, you know, when JNI came around, a lot of it was around contact centers.

Can I respond to my users? But in finance, we need that control, right? And the way I think about trust is it's no different than how we build trust with new employees. When we get a new human who joins our organization, our team, it takes a while to build the trust, right? Someone who's more experienced, we may build trust.

It may, you know, they may be able to demonstrate their capabilities a lot sooner. But a new analyst joining our team, we have to train them. We have to watch what they're doing. We have to look at the outputs, right? You know, and we have to provide feedback.

So I think about agents in a similar manner, right? Is we've got to give it that time. We've got to give it the time. Now, the challenges or the differences, humans work regular work days, right? Where you can see them, right? And if something goes wrong or if they have a problem, they can come and ask you the question.

Agents work 24-7, right? And there are more agents than what humans are. So the question is, how do you manage this large work? So it's almost like your spans and controls. Instead of having 10 people you're managing, it has now become 100 and thousands of people that you're managing, right? And so this is where you start putting in metrics and measures to see what is it that you need to monitor and confirm.

So I was talking with our tax team, and they have implemented an agentic solution on the taxes we pay. And I was like, how do you manage this? Because, again, this has regulatory compliance reporting requirements, right? You can't pay the wrong taxes.

And, again, so what is the mechanisms that need to be built in so that you can do your traditional closeout, right? You can check that A plus B from one system is equal to A plus B from the other system, right? So you've got to build those mechanisms in, and you've got to monitor them.

And, again, the good news is you can have other agents monitor them, right? So you can have an agent monitor some of these results. But it is – think of the trust as how do you build trust in your workforce, right? How do you monitor this? And is this just at scale? Excellent.

Thank you so much, Dhruv. So with that being said, let's move on to our final topic, the adoption roadmap. How to start, what to sequence, and effectively where to scale. And I know as a team, you know, we spend a lot of time discussing really the building blocks to agentic adoption.

And when we talk about this, we think about organizations having a modern data foundation, having all your data harmonized in context in one place with all the plumbing and everything for your data, acquiring it and as such all flushed out or to the best of your abilities.

Then we think of a, you know, really how you gather insights and having an analytics maturity. Are you consuming the data via Excel or spreadsheets? Have you moved on to a more, you know, aesthetic visualization approach or a chat approach or whatnot? Then we think about the governance and trust infrastructure as well that you, Dhruv, had discussed.

And finally, we think about moving to a position where you're effectively agentic AI ready, ready to deploy these new tools and solutions within your workspace. So with that in mind, Jose, one of the things we think about is that change management is where a lot of these programs stall.

You know, the technology works, but the people don't really adopt it. You know, what have you seen work, especially in finance teams where there's often a strong attachment to how things have always been done? Yeah, Perez, thank you.

Yeah, you hit it, right? Change management adoption is the key. It's the number one reason, number two, number three reason, right? Why these things go the way they do. I think Perez has covered, or Dhruv has covered a lot of this up front, right? The trust is a major aspect here, and understanding how the process will work and defining that process up front with the people that are going to be actually executing these functions in the future is key, right? Change management is sometimes seen as a training or a rollout at the end of an implementation or transformation, right? These new tools require that to be brought way to the front, where you're almost leading with the future state, the change management, the ways of working, and how everything will occur, and you're embedding the user experience up front.

So the people who will be doing the work, right, to Dhruv's point on, you're hiring a new agent as opposed to a new analyst, right? You need to be able to have those people bought in from the beginning, and not only at the working team level, but at the director level, the VP level, the CFO level, that this is the new way of working that we want so that there's support from the organization.

Because it's not just implementing a tool, right? There's talent, there's skills, there's other things that have to come into play as well. Awesome. So you mentioned people, and you mentioned talent. So that brings me to my next question.

Where does talent fit into this readiness picture? Are finance teams building the right skills internally, or is there still way too much dependence on IT and outside consultants? Yeah, so talent is an interesting element because it's front and center to what these programs have to be thinking about.

It's also something that is not just the role of finance, right? We're seeing that talent is an enterprise responsibility, right? So agentic AI and these tools are being embedded in the regular ways of working all of us, whether it's personal or business or operational work.

And so we're finding that the enterprise first has a role in establishing a foundational knowledge of training. I mean, we see that at Accenture with our own. We're all required to have a certain level of agentic AI, ML, whether we're deep into the technology or not.

We're all expected to use the tools in our day-to-day operations and bring tools to our clients and help our clients adopt those tools. So there's first an enterprise-wide need around talent. And then there's a specific role that each function, in this case finance, needs to play to ultimately understand how they adopt and how they train their people.

And I think a lot of that comes to the trust and the understanding, right? We need to explain how these tools work, right? We haven't gotten to a lot of the technical detail, but underlying the technology, it's not just a single program running and executing this, right? It's a layer of technology that is complementing each other to do this.

And so educating teams on how things work, give them the trust and that kind of an explainability factor on how things came together. And so that's probably the key. I'd say the last piece I'd say is that talent is an ongoing journey.

It requires continual learning and upskilling as the tools continue to evolve to your graphic at the beginning, right? If you haven't been learning along the way, well, there's some catch-up to do, right, to get to the next level.

Yeah. And perhaps I'll add one more thing here to what Jose said. It brings back to the FP&A teams. As we start using technology, whether it's AI, Gen AI, or now Genetic AI, the type of work people are doing is going to change, right? So going back to the narrative writing example, right, is our teams are not going to be writing the narratives.

They're going to be spending more time with business partners, right, working, doing some critical analysis. What are the skills required for hiring people in that position, right? And how do you measure their performance in these areas which may not have been as important or as critical in the past, right? So your metrics to measure performance and outcomes should also be reassessed and rethought of, right? So how do you hire for these new skills and roles where you have got to pay more attention on data analysis, business partnering? So that's another perspective on how talent should be thought through.

Well, fantastic answer, Dhruv, and thank you for providing the perspective. And that actually brings me to my last question for you. You know, when we talk about people and the leaders in the organization, what role should the CFO personally play in the adoption journey? And how is that different from the role of, let's say, an FP&A director or the finance systems team? Yeah, absolutely.

Now, the CFO as the leader of the finance function needs to be the evangelist. They need to be walking the talk, right? They need to be talking about it at all the meetings and giving real-world examples, right? Supporting their teams, their leadership teams, you know, giving them the right tools, the right incentives, the right capabilities on this journey.

They're the ones who are also providing them air cover, right? Because not all projects will succeed. Not all AI projects will succeed. Not all agentic AI projects will succeed. Or there will be unexpected delays, right? I can speak from example from some work that we've done ourselves.

It took us a little longer to get a data platform ready, which delayed the deployment. The CFO needs to provide that air cover and needs to provide the teams that support. The FP&A director, on the other hand, needs to, again, be very visible and be very, you know, show that they're leading by example.

So going back to that, providing for the global retailer, which I was talking about, is all these insights, all these comments were available in a dashboard. And what the FP&A director did there with his team was he said, when we go into these month-end meetings, I don't want you to bring me PowerPoint decks.

We are going to use this dashboard. Because what he was saying indirectly or directly was, yeah, we've spent a lot of time and effort. We've got a lot of collaboration. We've built this tool. We should be using this tool.

Of course. So that's the way, you know, when I say walk the talk, support the teams, encourage behavior. Those are some examples which, you know, will really make it different to the teams. Fantastic.

Thank you so much, Adruv, and Jose as well. And with that in mind, we're going to spend some time on Q&A. And some questions have already entered the Q&A. So we're not going to be able to get through to all of them.

And whichever ones we don't, please send them and keep entering them. And we'll get to them after the fact and do our best to respond accordingly. So I'm going to ask two questions. So number one, and whoever wants to take it can take it.

How do you make the case internally within FP&A for moving now? And especially when the technology is still evolving and the full ROI might not all be completely clear. Daruv, Jose, whoever wants to take that.

Could you repeat that question, Pras, please, once again? Yes. So how do you make the case internally now, especially when the technology within, you know, agentic AI is still evolving and the ROI might not be completely clear? So, you know, what can a finance team do now to take the initiative to, you know, lead this journey? Yes.

And this is, again, something that is coming up more and more. And like Jose mentioned, the technology is changing rapidly. You know, every day you wake up, there's a new capability, which either a Claude or a ChapGPT or even Oracle or even SAP or Workday or OneStream is announcing, right? So how do you keep track of that? So I think it's important to start with, again, the value equation, right? What is value to finance or to FP&A? Let's start with that.

Once we understand this, you know, we then start seeing how well we're going to unlock it. And this is where the two in the box becomes very important now, more so today than in the past, where finance has to work with the IT organization, right? Because you define the value and you say, okay, I want to write commentary.

I want to do variance analysis. I want to build a chatbot. Now, this is where you start having a conversation with the IT organization and your ecosystem partner. So if it is in the FP&A space with the OneStreams of the world, with the Anna Plans of the world and understand from that what is the roadmap and have a conversation with IT and say, okay, this technology is not ready.

It's going to be ready in six months. Can I build something today, right? And, you know, it's going to be partially through your work or we continue using it six months a year or do we wait six months a year? So this is where it becomes absolutely important to collaborate between the function and between finance and IT and whoever is your software provider.

Because more and more, I mean, I don't think we would recommend a company going out and, you know, building something if it's available on a platform, right? But this is where the companies need to make the decision is the way toward the tradeoff, right? What is the tradeoff between getting value today and getting value later? Fantastic.

And the last question I'm going to ask is, what's the most important thing a CFO should personally own in the AI governance conversation versus what they can delegate? Yeah, so, Prez, I can take that.

I think it comes back to the role that the CFO plays, right? So when it comes to governance and accountability and ensuring that the right people and humans are leading kind of those decision-making elements, I think is one.

I think the second is the CFO, like we've seen in the past, right? Even though it owns the finance function, is a key leader across the enterprise. And so ensuring that as a leader, they are ultimately driving enterprise transformation around things like data that are needed.

So that needs to be governed as well, right? The decisions made around data and source of truth and how that is fed into the finance org is important. We talked about the kind of the accountability, the trust, the explainability factor, and kind of how that leader can drive that in the org.

And I think the last thing is ultimately maybe not so much governance, but so much around how do we start the journey? Where are we going to prioritize initiatives? What is the value that we want? And ultimately coming back with, you know, what are the priorities, I think is the ultimate role that that individual needs to play.

Almost setting the vision, effectively. Yes. One of the questions that, you know, I've been thinking about, and I'm sure the audience has also been thinking about, is really what are the first steps to take and really a no regrets place to start when you're thinking about your agentic AI journey? Yeah, I'll take that one, Praz.

Where I think F&A teams especially should start or could start is with reporting. You know, teams are already reporting. They already are using some data for generating the reports, which is good enough quality because that's what's being presented to leadership.

I would start seeing as how can we embed AI, Gen AI, agentic AI into that process of creating those reports, creating the narratives around that, and, you know, begin on that journey. Because, again, you can start unlocking some value immediately there and freeing up time for your teams.

Fantastic. The next question is, well, more of an immediate step. What is something that a finance leader or CFO or director of finance, someone with influence, what is something they should do in the next 30 days to move along with this journey that is of critical importance? Yeah, so Praz, I can take that question.

I think it comes back to the readiness assessment we talked about earlier. Like, if you are not on the journey, you immediately need to get on the journey, right? And you need to understand what barrier, what gap, what maturity curve, whatever you want to call it.

But you need to figure out where you are on that journey. So you need to understand where you are with data readiness, right? If your data is lacking quality today, like, that needs to be addressed immediately with the various parts of the enterprise, operations, finance functions, whatever it may be.

Because whether it's agentic, ML, RPA, you know, working backwards through that technology curve, like, none of those things will work without quality data. Also ensuring clarity on the current state process and the people involved, right? As you look to automate and introduce agents, you need to really be clear on who is involved in the process today, what they do intimately.

So as you introduce agents, you are clear where those agents can play a role and how they can supplement the team. And ultimately, just asking yourself very clear, what is preventing you from enabling that next level and focusing on those building blocks that we've discussed today? Fantastic.

Well, Taruf, you have? Yeah, and I'll just add to something what Jose said. It's the classic pylon, right? But, you know, when we think about the process, so we should understand who's involved in the process.

We should also understand who's accountable for that process, right? Because going back to the comment I made earlier, right? And you asked the question, is human in the loop, human in the lead, right? Because we need to understand who's accountable because we have to ensure that whatever we are building delivers a good user experience for them.

Because we want them to use it because they're the ones who are going to be making the decisions based on whatever the agents do. So absolutely need to understand the person accountable for that process and people accountable for the process so that we ensure that humans stay in the lead and not an afterthought.

Fantastic. Well, thank you, Jose. Thank you, Dhruv. And thank you for your insights and the answers to these questions. And with that in mind, thank you to the audience as well for joining us today. And as I mentioned, if you do have any further questions, please submit them and we will do our best to get back to you as soon as possible.

And with that, I wish you all the best to everyone on your agentic AI journey.