Demo: AI-Powered Planning & Forecasting

00:33 - 00:47

Hello and welcome everyone. We're gonna give folks just another moment to get settled and logged in, and then we'll get.

00:47 - 01:05

started.

01:05 - 02:07

Alright. Looks like we have a great group dialed in.

Thank you so much for joining us for today's demo of AI powered planning and forecasting. My name is Najla Brown, and I'll be your moderator for today's session.

Before we kick off, just a few quick housekeeping items. First, today's session is being recorded, so you'll receive a copy of this presentation in your inbox.

On the right hand side of your screen, you'll see the q and a tab. Please submit your questions there at any time.

They'll come straight to myself and our speakers, and we'll address them at the end. Finally, you'll see a docs tab in that same panel.

I highly encourage you to check that out for direct links to resources you can save or explore later. And with that, let's dive right in.

I'm excited to introduce our speakers today. Om Kapoor is a global product expert at OneStream.

Joining him is Logan Shallenberger, director AI solution consultant at OneStream. Oh, Mike is all yours.

02:07 - 02:08

Thanks, Angela.

02:08 - 08:35

Everyone, thank you again for for joining today and, again, Angela for that intro. So really glad to have everyone here for the AI car planning and forecasting demo.

Just a couple of sort of housekeeping in terms of what we're gonna be going over over the next half hour. I'll take the first few minutes just to walk through where we see AI currently in finance and kinda talk about our approach embedding AI into the platform.

And then Logan will take it from there and show what that looks like in practice with the demo. So afterwards, we'll zoom out, talk a little bit about the bigger picture on OneStream.

We'll get back into your questions after that. So please keep dropping those questions in chat like Nigel was talking about.

We'll be monitoring those and, going through those later in the session. So let's get into it.

So, you know, I just wanted to start off this conversation by framing this discussion with two numbers. 63% of finance teams have fully deployed AI that's actively being used, and about 21% say that their AI has delivered a clear measurable value.

So that's almost two thirds of teams talking about having AI running in their practice today and about one fifth that can show real value. And that divide, that delta is really where AI and finance currently sits.

There's a lot of investment. Every team is being asked whether by their CFO or leadership about deploying AI and being able to deliver results that really live up to those expectations.

However, you know, those returns really haven't kept up. Pilots have really worked in smaller stages, but when you go to broad rollouts, that success is starting to stall.

So very few teams are able to really keep up the end, with value that they can actually measure. And so that really comes down to to two things.

Right? One one that we've all heard about is trust. Right? People don't fully trust the AI outputs, and there's a healthy skepticism around that.

Right? It's kind of a bit of a black opaque box sometimes. And, really, the the the that's less, to do with the AI tools themselves and really about being able to have the clear auditability and, you know, detail around how that AI comes to that particular output.

And the other is really around the fact that these tools are often disconnected from the governed data that finance uses. Right? And that that lack of business context to help explain and act on what comes out of that is really another key part of why, you know, teams aren't really able to drive forward and act on outputs of some of these AI tools.

So, you know, the real bottom line that we're gonna be talking about around is, you know, AI can't create value if it can't be trusted on, if it doesn't have the context it needs to execute on. So that'll be part of what we're gonna be answering as part of this next half hour.

So how do we actually act on and trust the AI, that your finance team's actually needed and what that looks like within OneStream? So with the approach that we've taken, you know, it's funny with the our monoclonal kind of talks about, kind of sums it up. Right? Sensible AI.

It's about AI that's sensible for finance, right, and built directly into, you know, OneStream rather than bolted on. And so the reason for that is OneStream is already the operating system for finance.

Right? Your teams already use it to run day to day, and it's all built on this unified data model and and govern data. So when you think a step back to think about how a business analyzes, right, going from analyzing to predicting and to then act on comes from having AI service within those workflows that teams are already running and using and are familiar with.

So we'll be getting into that in a bit with with Logan, but it starts with understanding what really drives performance. You know, what are the narratives to support it? Thinking about what are those real operational drivers that, you know, we talk about.

Right? Instead of benchmarking benchmarking against broad averages, you really go into talking about how business performs at at that deeper level and being able to understand what's performing well and where opportunities actually hide. And then that goes into the forecast with what Logan will show later on with Sensible AI forecast.

We're being able to pick up on demand, risks, and forecast shortfalls and being able to plan that out with confidence faster and more accurately without really having to lean on manual processes. Then finally, thinking about how do you actually act on that and actualize that value.

That's really with what we built with our Sensible AI agents, which were when GA back in May. And that really helped shorten the cycles and scale your team across hundreds of different agents to be able to do different types of analysis, you know, answering questions, writing commentary, and being able to do that all at scale.

The key part of all of this across both our quantitative and agentic offerings is really the fact that this all runs on a governed, auditable, and secure foundation on OneStream that your teams are already running. So you're not having to build a new stack, have data move back and forth, or have to, you know, train them on something new.

So one of the other things that, you know, we were talking off, right, kind of building off that agentic piece for a second. You know, agents are, you know, being able to scale across many different use cases and really help shorten that insight, you know, cycle from analyst analysis to insight to action.

And with what I said earlier, right, all that really comes down to being able to to work off of grounded data and the context that they need to understand, and and be able to realize all those insights. And so one of the things that we keep hearing finance team talk about is, you know, the idea of, you know, having agents anywhere.

Right? Whatever AI tools you're using, whether it be ChatGPT, Claw, Copilot, they want able they wanna be able to use those systems, and access the the the data, the financial govern data that they actually need to be able to have that necessary trust in the outputs that are coming out of that. And so one of the things that we had built as part of our agentic offerings is this finance agentic layer.

Really, the the core of this is to be able to bring OneStream, right, the system of record for the office of finance to these broader AI, tools so that you're able to have that financial, governed data at your fingertips so that all these agents are able to inherit that intelligence. Right? Our founder, Tom Shea, he always talks about how 80% accurate is 0% useful in finance.

That's really the kind of through line that we build a lot of our our solutions on. So you get that governance, whichever door you walk through, whether it's through our native AI agents our native agents or through any of these AI tools like your Gemini's, your Clogs, your Copilot's.

Right? So the semantic layer that we built is helps really understand which entity, which account, which year or scenario you're talking about and be able to really land on that right intersection. And so taking all of this with what we built with our Gendic offerings as well as our quantitative, offerings that we'll talk a little bit more about, in terms of what that means for planning and analysis here.

I'll take it off to Logan to kinda show what that looks like in practice.

08:35 - 24:59

Perfect. Thanks.

So let me go ahead and get my screen shared here. Perfect.

Now we're gonna dive into both our Sensible AI forecast offering and our Sensible AI agents, with the time that we have today. I think it's gonna be about twenty minutes worth of demonstration time.

Normally, this would take me about an hour, so we're gonna try to run through this a little bit quickly. So I'll hit the high level, points on each of these.

And then, if you guys do have interest in learning more, obviously, you can reach out to your OneStream sales rep, to dig into these further. Now we're gonna start with AI forecasting.

That's what we're looking at right now. You can kinda think about what we're looking at as an overview of a forecasting process.

One thing that I want you to pay attention to is this planning methodology column. AI forecasting is not intended to be a replacement for the entirety of a forecasting process.

It is intended to be a tool to focus in on some of the core pieces of your forecast that maybe are difficult to get right, or diff or take a lot of time ultimately to plan for those components, to introduce a little bit of automation around that and learn off of the historical data to be able to produce highly accurate and agile forecasts. Now the line item that we're gonna be focused primarily in on today is this guy up top with my AI daily volume forecast that's coming from my wholesale revenue account.

But I did at least wanna call out some of the differences of how this plan is coming together. Right? Even where we are using AI, we're using AI differently depending on which accounts you're looking at.

Like, for instance, these revenue accounts that you see here, we're using AI to plan at the p and l level for the account balance directly. Whereas the top line item that we're gonna be focusing on is a volume forecast on a daily level of detail, that is being used as an input into calculations to drive our revenue.

We also have tools available from the OneStream marketplace, like, for instance, our people planning tool down here, that ultimately don't necessarily need to leverage AI to get a very detailed and a very accurate plan, also incorporated in an efficient way. Right? So I just wanted to highlight this view to understand that, yes, we do have AI forecasting, and it is a very, very powerful tool.

But the beauty of OneStream is that we ultimately get to mix and match our methodologies depending on the nature and situation of any given plan that we're trying to forecast for. Right? Now within this wholesale revenue account, just to round out exactly how this particular account comes together, the output of my AI forecasting models is found here.

Right? We are forecasting by product, by region. How many units of any given product are we going to sell across any given region? That's what my forecast models are trying to predict for.

Obviously, we don't expect AI is always going to know everything, so we give end users that ability to adjust what's coming out of my models. That adjusted number is being used as an input into a calculation where we're bouncing it against our average selling price to calculate up to the wholesale revenue dollars.

Now, obviously, we can predict for just about any sort of number as a part of the forecast. Right? Like, in this case, we're predicting for volumes.

We could just as easily predict for the dollars associated to this revenue directly if we wanted to. My main question that I generally ask our customers is, are those units or are those volumes necessary or useful to other processes other than this revenue plan? Right? Are we gonna use these things for, components like labor planning, production planning, raw material purchasing, things of that nature? If so, then let's attack the units.

If not, and we're just using those units as an input into our revenue calculation, maybe we should just attack the revenue dollars directly. Right? It's the downstream implications of these numbers that really, I'll say, point us in the direction on which way we want to run.

At the end of the day, we can predict for just about any number that is quantifiable and measurable, in nature, right, and has historical data behind it. So whether that number is a dollar, a volume, or a unit, or even KPIs or metrics, all of those could theoretically be inbounds for our AI forecasting tool.

Now it's one thing to have an accurate number coming out of the models. On average, our customers experience about a 25% improvement in accuracy compared to their traditional ways of planning.

It's another thing entirely to be able to support and defend those numbers and to be able to explain why. What sort of impacts do all of the supplemental data, all of the variables that we're measuring against have on my forecasting output? Right? So like I said, it's one thing to be accurate.

It's another thing entirely to be able to explain why that accuracy is the way it is. I'm gonna skip a lot of the buildup of this particular workflow and just get right into the meat and potatoes of exactly what I'm talking about.

Inside of this tug of war tab, this is where we are taking our understanding of all of the different variables that we have loaded into the models. Some of these variables are intrinsic to the process.

Like, for instance, time is something that the system automatically accounts for. Time based relationships is something that the system automatically accounts for.

In addition to any variables that we have incorporated into these models as well. Right? That's the beauty of AI is it's very different than the traditional way of doing predictive analytics where you're looking primarily from a trend and statistical basis.

Now we're trying to understand the different drivers and variables that could have effects or, impacts to those trends more holistically. Right? So that's the goal of this model is to align all of those different variables that do carry an impact to your forecast into those models so that we can measure and understand the impact that they ultimately carry depending on which time period that we're looking at.

So this tug of war chart is effectively breaking down for us the math that the models used with the understanding of those variables to come out with the prediction that we did. What this is highlighting for me is that these are the components oops.

These are the components that affected my forecast positively. These are the components that affected my forecast negatively.

Right? What we can pretty definitively say is that because of the results from 365 ago, we've increased our forecast for the month of July by 58 units. Because of the day of the year variable, we've increased it by 45.

Because of the month that we're in, we've increased it by 42, so on and so forth. On the negative side of the equation, because of inflation, we actually decreased our forecast by 21 units.

Because of the year that we're in, this year must not be as good as, the prior years. We decreased it by an additional five units.

Right? This is how the models look at the world, taking the understanding of those relationships, the understanding of the impact that each of those variables carry, and applying it uniquely to this particular time period. And when I say uniquely to this time period, what I mean is that some of these variables can carry different effects depending on which time period you're ultimately looking at.

Like, for instance, for the month of March, the quarter is a reduction of our forecast by about 81 units. The U the month itself is a reduction by about 77 units and so on and so forth, whereas those variables were contributing positively to our July forecast.

Right? So the models go to a very low level of detail to try to understand the very discrete impact that each of these different variables carry depending on the time frame that you're looking at. Now the great thing about AI is that once the models understand that relationship that those variables carry, you can start to play with those variables.

Right? Scenario modeling. In this case here, I have two different scenarios.

I've trained my models on my pricing data. So I have my baseline forecast assuming that all of our pricing stays the same.

I also have a forecast here for what if we have a sale of 25% off. I've reduced our pricing in this case by 25% so that we can see what happens to our forecast if we go through that that promotion.

Now what we're ultimately modeling for is this number here. Right? How many units are we expecting to sell? Those units get translated into those dollars, this guy up here, based on that calculation that I showed earlier.

But, also, those units and those dollars are used to understand how that ultimately translates all the way down to the bottom line. Right? Any sort of, I'll say, relationships that are dependent on those units, things like cost of goods sold, for instance, are going to be carried out when we go through a scenario change.

So what you're going to see when I switch scenarios is, yes, we're gonna sell more units. But, ultimately, when we carry that impact all the way down to the bottom line, we want to understand, does selling more stuff actually help or hurt our organization? Right? So in this case, the baseline forecast are selling a 117,000 units, and that translates to about $9,200,000 worth of profit.

When we go through a sale of 25% off, what we're gonna see is that we're gonna sell more things. Right? We went from a 117,000 units to about a 121,000 units.

However, my wholesale revenue account was showing about $32,000,000 before. Now it's showing about 25.

And because of that revenue impact, now my profit for the month is showing about $6,700,000. Right? About a $5,000,000 profit hit because of the sale of 25% off.

My question in this scenario would be, do we have a lot of inventory sitting on the shelves? Are inventory carrying costs potentially fairly high? And is that going to be a potential $5,000,000 benefit or higher over the coming months? That's something that I can start to inquire into. Right? So this now gives finance users the ability to model for things that traditionally were very, very difficult to model around.

Right? Like, for instance, pricing changes. The what is aspect of AI.

Right? Now I wanna change gears a little bit here. I know that we don't have a ton of time remaining, so I'm gonna switch over to into our AI agents.

Our AI agents effectively turn all of our users into power users overnight. Right? That is I'm sorry.

Let me get my screen going here. This little button on the top right hand side of my screen pops open my agent panel and allows me to start interacting with the system using natural language.

I'm just gonna throw one of these suggestion questions in here in the essence of time just to highlight really quickly how this system works before we start to get into some of the automation that this system can provide. What the agent is doing when I drop this thought process down is it's trying to understand, number one, the nature and the specific details of the question that I prompted it with.

In this case, I said, show me my detailed income statement broken up by region. One of the first things that it's doing is it's looking to see if I have any standardized reports that can potentially inform AI on the structure of what I'm trying to build.

Right? I asked for a detailed income statement broken up by region. What it's doing is it's pulling my standardized p and l, this guy here.

It does not incorporate a regional view into that standardized p and l. So it's gonna take the structure of that report.

It's going to rebuild it with my regional dimension incorporated into that report. It's gonna run that report and then show me the output and try to start to reason across some of the data that was contained in that output.

Right? One of the things that you will notice with our AI capabilities is a heavy emphasis on transparency and auditability. Right? So up top, that thought process that I showed earlier, it's gonna walk you through all of the details of exactly what it's doing in a very non technical way so that you can always go back and audit and understand what exactly did the AI agent look at when it tried to get to these results.

Down below, when it starts to consume and show the outputs of those details, every single number from this report is gonna take you back to the report that the agent ran to surface this particular set of information. Right? That report is also visible down below in the source section.

You can see the structure of that report, the columns and the rows, even the point of view, the specific, dimensional segmentation that the agent cited upon, which queue, which entity, which time period, so on and so forth. When I open this report up, one of the things that I wanna highlight is that all of the financial intelligence from the system itself is also being abided by by the agent.

Right? Things like foreign currency translations, for instance, are automatically taken advantage of. Because this is a report, obviously, from a transparency perspective, you always have the flexibility to drill down on these numbers.

Like, for instance, this net revenue number that we're looking at here. If I wanted to understand which particular entities make that $400,000,000 number up, I can obviously drill down on an entity.

I can see the breakout of the dollars, the associated entities, and in this case, the associated local currency that that entity is denominated in. Right? Like I said, things like foreign currency translations are automatically taken advantage of because of the nature of how we attack the data by using reports as that mechanism, in between the agents themselves and the data.

Right? Because we're also using reports, one other thing that I wanna call out, you don't need to build a separate security layer to control the data, to control people's access to the data itself. Right? That is set up one time in the OneStream system, and it controls both the system itself and any AI access as well.

Right? Now from an automation standpoint, we've introduced a capability that we call analysis plans. Analysis plans allow you to use that same mechanism, but wrap it in a layer of automation around that process to say, I I have a process that I wanna support every single month.

You can set this analysis plan up to kick off on a time basis or attach it to different workflows or events in the system itself to kick off this analysis plan. Now it uses the same way of attacking the data.

It runs reports in the background. It consumes the outputs of those reports, and it tries to make, sense of what story the data is trying to tell us is.

Right? So for instance, in this income statement variance analysis, it's highlighting for me this column here of why each different line item matters and the and the respective variance in each of the different line items carry. In this case here, the income tax provision is one of the largest variance amounts that we have as a part of this process.

Now generally speaking, income tax provisions are are made, I'll say, at a high level, usually as a top line adjustment or a journal entry or something of that nature. So there's not a lot of detail behind it.

But this next variance here of about $600,000 for other OpEx, it decided it wanted to drill into further. The binary analyst, analysis plans are dynamic in nature in which they will drill down on the data automatically depending on the different thresholds and materiality triggers that we have set up.

Right? So in this case here, it's drilling down on gross margin. It's drilling down on cost of goods sold.

It's drilling down on OPEX. And then down below, it's also starting to pull into the conversation our different cost centers and things of that nature.

Right? It is smart enough to be able to navigate amongst the one stream, financial data that we have captured in our system, to be able to surface insights from that detail. Once again, heavy emphasis on auditability.

Down below, it's also highlighted for us the specific reports that this agent ran to get to this level of insights. Right? One last thing that I wanna touch on is also incorporating external AI into the OneStream system.

In this case, I'm using Claude. It is a very similar process of what we just ran.

It is a income statement variance analysis. You'll notice a lot of the same figures are also tying up to the report that we just ran.

Right? We have a high emphasis on, accuracy with these tools by nature of how they interact with the data itself. External tools also interact through mechanisms like reports in order to access the data similarly to what we were just looking at.

The one output that I wanted to focus in on is actually at the end of this conversation. It took all of the data from this entire income statement analysis and it threw it into a reporting package in PowerPoint deck for me on the fly.

So any processes that are potentially dependent on OneStream data downstream and of OneStream itself. Like, for instance, back in my time in finance, we used to go through cost center review meetings every single month.

It would take our finance team an entire week to break down, all of the presentation packages that we need to put together for each given cost center review package. Right? What if we could automate that entire thing? That is totally inbounds now with some of these external tools being plugged directly into the OneStream system, having a high degree of confidence that the data is coming out accurately and correct.

All of this stuff can start to be automated. Like, for instance, in this case, this reporting package for my income statement variance analysis.

Now with that, I wanna make sure that I give Ohm a little bit of time to wrap us up here and talk about, OneStream a little bit. So, I'll be around, for any questions that you might have at the end.

But at this time, I'm gonna go ahead and pass it back off to Owen.

24:59 - 27:59

So really appreciate it again, Logan, just kinda going through that demo, you know, really giving us a a really good, a view into what that looks like, especially with, you know, with since we're forecasting with agents. I know everyone's been excited and kind of talking through, what we're building out on the the agent side here.

So just to kinda wrap up, just a a very high level around, you know, the the OneStream story and vision from a from a broader perspective. Right? We we kinda talk through what it looks like specifically on planning and forecasting, but I think the the the broader story of what I mentioned earlier.

Right? The idea of us, you know, our our vision is really to be the operating system for modern finance. And, you know, that really comes into play with us being uniquely unified, infinitely extensible, and and AI powered as to what we just showed here.

And so what that really means is having that one platform and one data model. Right? So it's not just a point solution or a suite of different solutions that are packaged together.

Everything runs and works off of the the same governed data. So the same numbers, there's no reconciling between different systems.

You can, you know, really stand behind everything in terms of, the the outputs. And so, you know, being unified in the middle of having that financial operational data together, that's really where everything gets kinda gets built out.

So and we start off with, you know, you must do your core finance work. Right? So that's your close, your consolidation, your reporting, your your planning analysis.

And with that, right, all that planning work, you know, loves to to guide some of the operational work like we're talking about earlier. Right? So that's your workforce, your sales, your operational data at any level in-depth that you need to.

And that really comes into the part with what we talk about with extensibility. And with all of that said, you know, the AI is just embedded directly in.

And so with our portfolio, as, Logan kinda showed earlier, it's all built into the platform in those workflows that teams are already using. Right? So you're not having to to move data around or have a different tech stack for that.

And so we also open that up with what, Logan showed earlier with Claude with the plug and play architecture. We bring that out, the agentic layer to your Claude, your Chachapatis, your Copilots of the world so that you're able to bring that financial intelligence into the tools that the finance are already working together.

So, you know, working with OneStream, you can really bring finance further. And that's kind of the the main ethos for, you know, what we've been kind of running through today and talking about with AI powered planning and forecasting.

So, I know we we threw a lot at that at you guys all within a thirty minute stand here. So we'll take a a bit of moment just to kinda go through some q and a and and kinda look at some of the questions that, you know, we had kinda pulled in from the chat here.

So, Logan, one of the the big questions that we're getting that everyone's really excited about is, thinking about as people start using agents, start thinking about, putting that out in, in kinda in application. We talked a little bit about governance and audibility, but can we kind of just take a step back to kind of talk about how, you know, an admin can actually go through and actually monitor what that looks like from from their.

27:59 - 29:23

perspective? Absolutely. Yeah.

So here at OneStream, we've always had a, I'll say, a a hyperfocus on auditability across the system. Right? Every interaction that a user has within the system is tracked forever.

It's not something that you can turn off in OneStream system. We take that same point of view with all of our AI capabilities as well.

Like, for instance, with our AI forecasting tool, every single iteration of a model run itself is saved forever. You can always look at the training processes, the different impact that the variables carried at different points of the year when you went through those training processes, so on and so forth.

Same with our agents as well. Every single interaction that any end user carries with the agents is also saved forever.

So there's administrative tools. We call it our AI control power that ultimately gives the administrators of of the OneStream application, control and authority over those agent driven processes, right, both inside the system.

So any interactions like you saw I had, but also external to the system as well. Any calls into the system from, an external agent like Cloud or Copilot or what have you is also going to be tracked and visible for your administrative team.

Right? And it's also saved forever. So we also take that very, very heavy emphasis on auditability with everything that we do, respective to AI.

Perfect. Yeah.

And and I think that one of the the core things that.

29:23 - 29:54

we've been seeing in this space is is having an eye for governance and auditability across the system. So that that makes perfect sense with kind of the way that we've been kinda seeing having that that background and and and the eye for for governance there.

Another question here, specifically around, sensibly eye forecast. How how involved, does the finance team have to be with a data science team to spin that up and, you know, in terms of maintaining and and and kind of deploying that out into the field?

29:54 - 30:20

Great question. In my experience, data science teams, have been extremely useful as part of those projects, but are not required by any means.

Right? We have within our both within OneStream and within our partner ecosystem, very, very powerful teams, very strong teams that can ultimately implement and support, AI forecasting projects and initiatives without the need of data science. Right?

30:20 - 30:21

So I I like to.

30:21 - 31:11

think about it like, you know, you guys bring the financial intelligence. You guys bring, all of the insights into your business on the the nature and types of things that could be affecting your business, and we will translate that into, I'll say, the data science activities that are needed to support that.

Also, a lot of those data science activities, we have built automation around. Right? That's why our AI forecasting engine is so powerful.

If we've taken best practices from the data science community and ultimately automated a lot of those best practices to our engine. So all you really have to do is bring the data that you want to train the models around, maybe make a couple of tweaks to that data.

We use anomaly detectors to cleanse and sanitize the data before we get into training, and let our engine effectively take over and do all the work. On an ongoing basis, no data scientists needed.

Perfect. Yeah.

And and I think one of the things that.

31:11 - 31:58

comes with that too is in terms of the maintenance of these, different, applications, that's all done within the OneStream platform itself. And so, you know, it's fully financed owned.

Right? So you can bring a data science team to your point if you need to in terms of the evaluation kind of building out that initial, lead into for the projects. But at the end of the day, it's all, you know, built for for, the office of the CFO.

Last question here. You know, we we kinda went through a lot right now.

We went through forecast. We went through agents.

There's a ton more kind of across the fold here. But, you know, when we when we have the conversation around AI, how do you really, see teams starting off and kind of building out either COE a around AI or or thinking about implementing that as part of their their day to day?

31:58 - 33:34

You know, from all the experience that I've had, I've been working in AI here at OneStream now for the better part of about three years. It's a little bit of a different process for people to incorporate AI than, Allstate traditional, software was, where, traditionally, people had Allstate processes that are in place today that they're potentially looking to transform or put a system around or put a little bit of automation around or what have you.

Generally speaking, not a lot of people are doing a ton with AI or have not been doing doing a ton with AI up until probably the past year or two. So we like to come into, the conversation, I'll say, with open eyes.

Right? Because most of the time, a lot of the stuff is not new to a lot of folks. And for most people, I think that from, like, an ease of adoption standpoint, it's much easier to adopt AI as a part of a system that you already have in place as opposed to trying to build something custom unique to your organization.

Right? It's no different than any other software. The build side of software is only one side of it.

Right? It's also the administration and the maintenance and the upgrade, support, and everything on an ongoing basis that is also very, very important. So adopting AI as a part of the systems that you already have in place today, is probably the easiest path forward.

But like I said, it's you kinda gotta approach this with a little bit of of open eyes on, you know, what's possible and to really dig into kind of that part of the possible. 100%.

And I think that that, really leads to, a.

33:34 - 34:23

a great point where, you know, as any of these questions kind of pop up for, you know, people on the on the webinar right now, you know, reach out to your OneStream teams. You know, we're we're more than happy to kind of go into kind of talking about the art of the possible and thinking about how we can, you know, help, provide that value and think about how, you know, each of these different use cases, kinda plays in in part for your particular business and processes.

So, you know, with that, you know, really appreciate everyone joining us here for this half hour. You know? And if you have any questions going forward, you know, please reach out to, your OneStream team, and and we'll be reaching out, after this individually with with an email.

So thank you guys for for joining us here today. Thank you.