Neil Thomas 0:00:13.8: 

Good afternoon everyone. Welcome to Finance Innovation in AI and Apps with, the highlight will be my friend Becca who's the only person in the front row, thank you Becca! To give us some real world lessons from Spaulding Ridge, about how to leverage all of this innovation. I'll thank you for spending time with us, I hope it's warm enough in here! That big room is really, really cold. So my job is to warm you up, get you all excited about AI and innovation, and applications for Anaplan. Okay, so my name's Neil Thomas, I look after all of our application portfolio from a go-to-market perspective, so I'm the kind of person that works with the team, who build the applications, bring them to market, work with ourselves, team, our partners, our implementors and hopefully bring them to you. If you are not on the application train you should seriously consider getting on it because they are fabulous, and I'll be delighted to spend more time with you. 

 

Neil Thomas 0:01:14.0: 

I'd also like to advertise at this point I think my friend EJ who I work for mentioned this, this morning, but we like to collaborate with our customers on how we build these applications. So if you see an application you're particularly interested in, or have a brilliant idea for an application that you think should be appropriate for our wider customer base, please, please come and talk to me, and we'd love to have as part of that influence group, participate as much or as little as you would like. The applications are built with the customers in mind, and with your input they'll be even better for everybody. So here's our agenda, we're going to do a little bit about AI adoption in the world today, and bring it back to practical use cases with Anaplan and some of the content that you saw this morning. We're then going to do a little highlight of some of the applications and the kind of journey that you may go on. I'll do that pretty quick, so that we don't exhaust you, and I’m going to tell you why finance is in the best place to leverage this innovation in certain different roles as well, so appropriate to your culture, your company, your ability to invest. How can you take advantage of this? Then we'll bring up Becca for about a 20 minute conversation and hopefully some of you will ask questions at the end and we'll have a great time together. 

 

Neil Thomas 0:02:26.6: 

Okay, so let's talk about AI exploration, everybody's exploring AI, you may notice I have a little grey in my beard, only a touch, I've only just started showing some grey. I've been around in the planning space since 1994, so yes I am truly that old, I know I don't look old. These sorts of challenges in the marketplace haven't really changed that much, other than they accelerate. So the chaos gets faster and more swing, the demand of data and challenges and integrating just get bigger and bigger and bigger, and of course we now have this new catalyst event that's happening to us, which is what do we do about AI, how do we really exploit this new technology. So against that background, what kind of adoption curve can I think of from an AI perspective. We're all at different levels of maturity, I did a recent event, roundtable with about 40 people and literally not one person in the room had started an AI innovation project in finance! So that was a very immature audience who were just there to learn about where might they start. Another event there were other customers and prospects who were much further along and had already started to adopt internally. So we're all at various stages of this maturity curve, and the good news about our innovation is you jump on at the right spot based on your knowledge and expertise. 

 

Neil Thomas 0:03:49.4: 

Exploration, most of us at that stage, and when you look at a Gartner report, 2026, so pretty recent, survey of about 150 CFOs, that 60 per cent of those organizations are in this exploration stage. So if you haven't started, you clearly not alone, obviously 40 per cent of the result is we haven't started yet, we're still thinking about it, etc. It's not a low cost innovation either, you need to get return on this investment, you can't just go off and play, you need to get a real value back, so that's why customers have obviously been very hesitant, very thoughtful, very sensitive about this before they get on the train. The really bad news about AI is same survey 83 per cent of our CFOs are still waiting for results to be seen. So over the last 12 months that they've been making these investments and learning, well where is the innovation, is it truly giving me the value that I expect for the dollar that I’m putting in, and obviously we are very conscious of that at Anaplan. How do we actually give you AI that's actually useful. The most interesting thing about what prevents ROI are these four factors. So clearly identified, the biggest one of course is data, AI is only going to be good as the data that you put in the system. Now we've heard that about everything, my reporting, my analysis, the value to my end users, it's only as good as the data. AI is just another end user of our system, so we've got to get our data right, I know Becca's going to talk about that a lot. 

 

Neil Thomas 0:05:19.1: 

Culture wise, we've got to have the right culture. So if I'm an innovative culture and I lean in and I like the new stuff and I like playing, I'm going to have much more success. If I am a skeptical organization, or I don't like change, you're almost guaranteed not to have success, so you've got to lean into this new technology, we're going to be bumpy, we're going to learn a lot together, but we've got to have the right culture. Got the pick the right process, is it going to add value to this particular process, is the process perhaps an easy process to start with so I can immediately get value, so pick the right process. Of course security, I've got to make sure it's compliant, that's one inhibitor of actually getting started of course, is it going to be safe, is it going to be in my environment, what am I sharing with the outside world and of course you don't want to share anything. So those are the four inhibitors, but primarily data and culture are the biggest ones. Same survey, what you've got to do is get the finance data right, and we've heard that forever, it's about as old as me, and as old as the first slide we looked at, it's a constant problem of getting data into the right place to give context and let the AI go to work. When I think about our AI and how it fits to this cycle of curve, this is where you can decide how you would like to jump in. 

 

Neil Thomas 0:06:31.2: 

So on the left hand side if I'm exploring and I want to keep it within my finance team, you saw CoModeler today, that's the perfect starting point. How to document my applications, how to extend my applications, how to build my applications more quickly, it's great and it's just within the finance world. My end users don't need any AI experience to do that, we don't change the UI, that's all a core building facility. Really great place to start from an exploration perspective. Possibly the highest value, real time savings, very tangible on how fast CoModeler can do things for you. When you get to the middle stage you start talking about planners and analysts, so these are people who use what's been given to me in a model, they may be a finance expert supporting the business, they may be an end user, and they could start using things like an analyst agent to go off and answer questions, make recommendations, those types of things. That's where culture becomes very important, and obviously we've got to get the data absolutely right for it to be valuable. So it's probably a heavier lift, but it starts to affect all of our end users, and we're not quite there yet as Joe Horsey talked about in terms of our virtual assistants and our agentic framework, but that's coming very, very quickly, where we start connecting all of these analysts and co-modelers together for really, really high value. So we can kind of hold that one, but we're getting there. We can't do the last one without the first two. 

 

Neil Thomas 0:07:55.3: 

So when I think about this and trying to address the problems that we solve, we can think about if I've got my data pretty well organized and I've got my expert model builders and we've got the right culture, CoModeler is a great place to start for my model builders. If I've got the right process and the culture now that my data's ready I can extend this to my end users, get really good feedback on how to improve, what questions to ask, how to train the AI, and really start to add value. Of course if you've seen any of the security compliance framework we have, we are pretty much compliant in terms of any latest, greatest standard as you would expect, particularly with our customer base which includes a hell of a lot of financial services who are very demanding in this area, so making sure we have the right security both in terms of the data and the way that AI interacts in terms of your tenant and your applications. So that's how we can go on this journey with you, again you can jump on at the right spot and then mature slowly but surely into the right direction for you. As a result of that we do think that you'll start to tick these boxes that prevents success, and again Becca's going to bring some real world examples to us around how to do this in a controlled manner so we don't go too crazy, we don't spend too much money, too much time, get our wins and then progress from there. 

 

Neil Thomas 0:09:10.1: 

Let's talk a little bit about applications. So what enables AI to work is having the data in the right place of course and the right context and structures, and this is part of the reason that Anaplan has built applications. We can train for you, if you build your own application, not only do you need to train the AI but you need to test it and bring it to life, we do that for you with the application framework. So we believe this is a great way of unlocking speed, time to value, I start with a framework that's already there, I can use all of my brilliant Anaplan skills that I already have to extend it, add value where it's needed in my process, add my new other models, reconnect it to other things, but I get that accelerated starting point and I get innovation. As Anaplan comes out with innovation, we load them into our applications, you upgrade your application, you take immediate advantage of those innovations, you no longer have to do that case-by-case, customer-by-customer, you can follow us along. So very, very important and thoughtful roadmap. I like finance, I'm a CFO from a previous life, so I like finance, and I like finance as an area for this, partly because finance sits at the hub of everything, all roads lead to finance, finance is the most important part of the company of course, keeping all the numbers, working out where we're going to invest our next dollar. When we think about AI, we're thinking about how to help three key areas, and when we think about our applications we build our applications in this way as well. 

 

Neil Thomas 0:10:39.4: 

So we have, off to the left we have our CFO, sits at the top, would like a framework, one consistent place, one vendor, much easier to work with, obviously with Anaplan's enterprise wide planning vision we bring that to the table. So we'll keep her happy. In the middle we have our FP&A team, and we're going to dive a little bit deeper into FP&A, but basically FP&A need the flexibility to build the planning models that they need to run the business, and everybody does it differently, it doesn't matter if you're one retailer versus your nearest competitor, you'll have a different culture and a different process and a different philosophy. So we need that flexibility and planning is very unique in that way and obviously that's a core strength of Anaplan. However if you build everything unique and from scratch it can take a long time, why not provide the framework of the applications which gets you 70 or 80 per cent of the way there, keep upgrading for you and then you extend the value to those unique areas where it's most high return. On the right hand I'm sure some of you are aware, we provide a consolidation product, very originally named Financial Consolidation and Reporting. Great consolidation product, extremely user friendly, very, very modern, integrates beautifully with our FP&A solution, but we do keep them distinctly separate. We integrate, but the processes are different. My audience for a consolidation process is a regulator, it's a statutory compliant thing, it's got to have structure, it's got to be auditable. FP&A can be a bit like the wild west, we're constantly creating new scenarios, constantly changing things, we don't do that to our consolidation, but we need the two things to talk together and that's how we've approached this particular problem, all within again the same Anaplan framework. 

 

Neil Thomas 0:12:21.4: 

EJ showed this slide and you can see how we think about this calculation layer, we have a consolidation calculation, we have a planning consolidation, we then bring AI calculation into play as well. Data transformation glues everything together at the level that you wish it to be glued together, so we have customers that use our consolidation system but never actually take those actuals and put them in planning or vice-versa, the two things are completely siloed. We have other customers where they are dynamically real time essentially integrated, so again you can configure using our data orchestrator any way that you want that level of integration. Sitting on top of that of course you have our applications, or the applications that you have built, our applications are built just like you would build them with some special sauce underneath. Then last but not least we layer on the AI, so as we've built on our calculations we've got the right data, we've got the right context with our applications, we then add and train the AI agents on the things that we've created. Again we're doing that constantly, and we are the early innovators in the application team of terms of being that customer, we're the alpha customers, we make it all works, we make sure it works and then we bring it to you guys in the market. 

 

Neil Thomas 0:13:34.4: 

FP&A is particularly interesting because essentially there are three types of FP&A, our friends in corporate accounting are going off and doing the same thing, same clones time-after-time, the trends to have FP&A more and more embedded in the business again suits this Anaplan model. We have corporate FP&A sit at the top, M&A, strategy planning, those types of activities, they require certain applications and certain models and certain views. We need the same data for that to drive down into our business unit, maybe it's a manufacturing business unit, maybe it's distribution, still need their unique processes. Off to the right we now have obviously functional FP&A, that's the FP&A that's embedded in, whether it's sales, HR, operations, and they act as almost the CFO for that particular team. I work in the development group at Anaplan, we have essentially a functional CFO that works with us making sure we're spending money on the right things, all she thinks about is how to support our function and she reports into the corporate CFO. So when we think about our applications and our structures this is how we're thinking about it, and we try to make them as common as we possibly can, and again getting you a certain percentage of the way there so that you can customize and configure from there. All three areas can benefit. 

 

Neil Thomas 0:14:50.1: 

One of our favorite applications in operational workforce planning, absolutely perfect for HR, where I have an HR function who has an HR FP&A business partner, who collaborate together, and it brings the best of both worlds. In my day the most annoying question will always be, how many headcounts do I have, and HR would say, 'We've got this many,' and finance would say, 'We've got this many,' and payroll may actually say something else. Operational workforce planning reconciles headcount, like the biggest, the most immediate benefit, and then flows into the operational model that you have in FP&A. So we believe our applications suit each of these audiences, whether it's a business, a function, or at the corporate layer. How would you go on a journey with an application? So you can do this in any order that you like, this is just an example. So maybe I'll start off with consolidations. So today customers replacing some of their aging consolidation systems, some folks still do it on Excel, but we'll start with consolidation. 

 

Neil Thomas 0:15:50.4: 

So I need to close my books, consolidate my numbers, get my actuals, report them out to my regulators and my management team, etc. Great, now I need to do some planning, I'm in the corporate FP&A team, maybe my business unit, FP&A team, now I need to start doing my hardcore planning processes, drive my next forecast, my next cycle. We have applications that can help that depending on the domain that you're interested in, all built around the core application of integrated financial planning, the three statements essentially, the framework for those three statements. We might be a company that requires supply chain, so we can start to integrate into the supply chain and to extend into that area, if I already have a supply solution that I really like, that's okay, it's just a data source, we can turn those hexagons off and we just use whatever you have already. Similarly I can move off into my other business unit areas, such as obviously HR what we talked about, or revenue performance management on the right hand side, particularly focused on large sales teams, thinking about how they go-to-market, allocating quotas, territories, etc. Again you can pick and choose what's appropriate, when you look at an application that's a gap you can create your own application or take the best of what we've done, look at it, steal all the brilliant ideas and then implement that yourselves, it's entirely up to you. 

 

Neil Thomas 0:17:08.3: 

That's what a journey can look like across enterprise wide planning, and we have an ever increasing amount of customers that are starting to complete their journeys here and obviously you've heard about [?and video 0:17:19.0] is a good example of that this morning. When I think about Silicon Valley, I think about SaaS companies, why wouldn't I, so software, so this would be a package for a software company that's a business SaaS company, thinking about annual contracts, multi-year deals, that kind of thing. This is the perfect solution for them, if they implement, and they have implementation services, project cost and project resource planning would be appropriate, if they don't, just pull those out and just worry about selling software and let the partners do the implementations, or perhaps consumers implement on their own. So you can obviously put these things in different combinations, and when you look at the framework, the underlying data orchestrator, obviously we are moving these data around whatever applications that you may have, sitting on top. So I don't want to say it's easy, I don't want to say it's seamless, but it's very, very packaged, very quick, and obviously if you take the vanilla application, let it run for a while, find the gaps, find the things that you really want to fine tune, that is the perfect implementation approach. Versus going to the whiteboard and then deciding to design everything originally. We've probably thought about again 70, 80 per cent of the standard processes, given our customer base is very large and we have customers help us with this, you can take the advantage of all that and then move on from there. 

 

Neil Thomas 0:18:39.6: 

If you've already built some of this and you're thinking about, I really want to start with that project stuff, I've got my P&L, I've got my RPM stuff sorted out, that's great, you can just bolt those on not a problem. So that's just an example of a SaaS company and obviously we have them for retail, we have them for manufacturing, any other examples. So when I think about Anaplan applications, to me they bring together all of the built in expertise that we've collected over the years, so 2700 or so customers, we know exactly what they built, how they built in, we know the wins and the losses, the hard things and the difficult things. We've put those into the applications to make sure that our customers get a best practice leverage, the right structures. We've also built it on scalable future ready architecture, everything we build is on the latest, greatest stuff. When we go test new features, we test them with our applications, so of course it's ADO compliant, of course they're built on Polaris, of course they have our integrated analyst agents, of course we've tested it with CoModeler, etc., etc. 

 

Neil Thomas 0:19:43.7: 

Then obviously in the middle yes, this is where we're building all of this beautiful data together, all of these best practices together, and then able to leverage the AI on top of that, so we've got the data in the right context, and so now it's going to be valuable. So that's a little bit about AI and that journey, and fitting to where you might start, and a little bit about the applications themselves, and obviously there's plenty of demos going on as well so please take a look. With that of course how do you do this in the real world? So Neil stands up, he talks about this, all sounds so good, I have an English accent, if everybody is American in the room you all trust me immediately because of my accent, right. So how do we bring that to life, what does it look like in the real world. So I'm going to ask… Oh no, I'm not, one second, Becca, I forgot something. I meant to advertise why finance is the best team for this, so again I'm biased, I love finance. You guys know the data better than others, the most important thing is you know the bad data, and if you're not going to fix the bad data problem AI is not going to fix that for you, so focus on where the right data is. The second thing is the process, you know the planning process, you know the things that are repeatable, the things that AI can truly learn from, whether it's in the predictive area, have I got enough data to run my predictive analytics, or whether it's just in the routine agentic, automation, add value, speed things up, remove my manual work type of activity. Whether it's using CoModeler, you know the process. 

 

Neil Thomas 0:21:13.9: 

You also understand your users, are they ready to use this, if they are not ready to use this do not invest any time in it, wait for them to be ready, Or perhaps pick a pilot group, spend a little bit time with Nasdaq in our New York Connect, they picked a pilot team, they built what they wanted to build, and then they showed the other teams who were very, very skeptical, and those guys went, 'Yes, I would like to adopt this,' so pick that pilot, the right users with the right culture, and again obviously make sure that we're security compliant as possible. All right, so let's move on, sorry about that Becca I forgot to do that bit! All right, so how do we do this in the real world, so I'm delighted to have Becca join me. Becca Buell from Spaulding Ridge, so she's the brains of the outfit of course, she's the one, the expert in Anaplan implementations. I'm going to ask her to start and tell us a little bit about yourself, and then we'll start to bring it to life for the audience. 

 

Becca Buell 0:22:08.6: 

Sounds great, thanks, Neil. For those of you who I haven't had a chance to meet yet today, I am Becca Buell, I'm a managing director at Spaulding Ridge. I'm based out of Chicago, but I’m often here in the bay area, originally from Green Bay, Wisconsin, so go Packers. Prior to Spaulding Ridge I did work at a SaaS tech company, Twilio, and our organization at Spaulding Ridge really focused on the office of the CFO and how to help support finance transformation. It's often at the intersection of a lot of the things Neil's talked about which is data, process, people, and technology. I specifically focus on supporting high tech and media organizations, and peers of mine help with CPG manufacturing that are dealing with similar problems just with different drivers. Companies that I've worked with, some of which are some of the largest media organizations out there, you've probably watched some of their shows on your couch this past week, or this weekend while you've been travelling. Also some PE backed tech companies that your organizations also may be using to help with your spend management, so excited to talk to you about some of those learnings. 

 

Neil Thomas 0:23:21.7: 

Cool, can you just give us a couple of your favorite projects? 

 

Becca Buell 0:23:25.5: 

Yes. 

 

Neil Thomas 0:23:25.6: 

Forget all this AI stuff, app stuff, just tell us a couple of your favorite stories about people adopting Anaplan. 

 

Becca Buell 0:23:31.0: 

Absolutely. I would say one organization we helped with their go-to-market planning we got a process that took 90 plus days down to under 30 days, so we got to give some really hard working folks time back to either reinvest in their own education, and like ways that they can be a business partner, and also help the leadership team with that end-to-end. Then a very large workforce transformation that from a finance transformation we started with the foundations and now we're starting to infuse with AI use cases with Anaplan which I'm personally really excited about because I know a lot of folks here are trying to figure out how to best take advantage of those use cases. 

 

Neil Thomas 0:24:10.7: 

Yes, I'm glad you mentioned that because one of the things you and I talked about was the difference between finance transformation, and then AI transformation. You can't have the second without the first. So any thoughts on just when am I ready for finance transformation! What should I do first, how do I prepare for AI? 

 

Becca Buell 0:24:28.6: 

Good question. I think AI has been a huge hype over the past year, two years, and I think that finance is in a point, especially in high tech and media where there's a lot of change that's being put in front of folks, there's demand signals, customer changes, like pivots and finance is expected to do a lot more, move faster, but oh by the way let's keep headcount the same, which how do we do that? So AI has become a forcing function, not because it's a hype, but because they need to, we all need to take advantage of it to do more. So what I've seen on a maturity curve is we look at our organizations from an automation intelligence, generative and automation, like agentic automation, and rather than running directly to full blown autonomous agents, it's better to look at the organizational readiness and where you can start with potentially automation or intelligence. That's standard finance transformation, that's not necessarily always AI. 

 

Neil Thomas 0:25:32.2: 

Yes. So when you get engaged on a new project, how do you evaluate if the customer has the right data, the right process, this is the right choice, higher guarantee of success, risk mitigation? 

 

Becca Buell 0:25:45.8: 

Good question. Before we even look at data, and process, and people, the first question I usually ask either the head of finance transformation, the CFO, whoever I'm working with, is what part of the P&L needs the highest amount of change, or what do you have the least visibility into, what needs to improve the most. Then on top of that, what operational metric will actually drive that P&L line item. So there's one CFO who is talking with me about how he needed to materially improve NRR, an in order to do that he needed… It wasn't just automation, he needed help with intelligent AI to help with informing early indicators of customer behavior, to change that. Another organization needed help with workforce management, and the spend, because the higher dates change so materially as the year went on. So they needed automation and then also intelligence on baseline hire dates. Another organization needed help with board deck creation, I know those that are responsible for those, they're very painful and arduous to put together, so creating a baseline and picking from that before, once you find that item or outcome that needs the most change, then double-clicking and looking at is my data ready, if it's not probably not the best place to start. Are the people who use this, and consume this metric, AI forward and open enough where they trust and make decisions off of that. Probably not going to happen the first forecast, but would they at least be open in maybe the second or third forecast cycle. 

 

Neil Thomas 0:27:21.6: 

When you think about a project how… So if I'm in the audience and I go, I think I've got the right process, I think I have the right data, I have the right attitude, I have the right culture, we're going to make a change. What's the timeframe where of course we're all terribly impatient, so the timeframe from have the idea, call you, work through the process to, yes, this is working. How do you get early indicators of success, how would you measure that? 

 

Becca Buell 0:27:48.8: 

Sure, so one thing that our team's done is an AI lab that helps with whether it's a two day workshop, or a four to six week project to help identify, part of that framework is to set what the baseline today outcome is, so if it's today it takes me seven days to produce this deck, and we want to get down to a hypothesis of at least shorten it down to three days. So as a part of this lab we have to create a framework for what the current state and what you eventually want it to get to, unlikely that that's going to happen the first time you use the new product, because there's change management and trust and all of that. So some early indicators of that is are we able to produce it faster, is there more response time, are folks spending less time debating the data, and more debating the assumptions that contribute to the model. Those sorts of like more organization and behavioral indicators. 

 

Neil Thomas 0:28:46.6: 

As I'm an application guy, one of the things we advocate for is take the application out of the box, get used to the gaps, and then fill the gaps that are worth filling. The other ones standardize, use the same process, you'll be fine. How do you see that working in this process as well, because it's a little different from two years ago where it's a whiteboard and we get started. Have you seen applications help with that, or resistance, or both? 

 

Becca Buell 0:29:15.9: 

Yes, I think that where AI is best served, or where it fits best is where there's a clear process that's really well defined, there's stable structures, there's folks that are leveraging the same consistent dashboards for example, and I think the apps serve well to that framework because there are structures and data elements that are common from a data model and also people and governance that AI can be built on top of. So I've seen a little bit more acceleration and value from AI, whether it's with an app or a more structured data model in general. 

 

Neil Thomas 0:29:49.1: 

Got you. When you walk in and meet folks, how do you know if they're ready, like the personal bit, do you have a criteria that we need to be aware of? 

 

Becca Buell 0:29:57.2: 

That's a good question. I don't know if there's a specific persona, or personality that's going to be open to it, I think what I look for is if it's one person that's coming to me and saying, 'Hey, I need to infuse AI into my FP&A team.' Do you have the right planners that are interested and actually want to change what they're doing. Do you have your enterprise AI governance from IT or other parts of the business that are defining that involved. Do you have the data team bought into to help make those updates? So it's very organizational, less on one person, but you do need… The areas where I have seen AI be most successful is there's somebody who's very passionate about rocking the boat and making changes, and it can't be passively, we might do AI, we might not. You need to be all in, because it takes everybody following behind the leader, and I've seen a few, if anybody's looking to talk with a really strong AI transformation leader, I have many that are here based in the bay area and other parts of the country I'd love to connect you with them. 

 

Neil Thomas 0:31:01.8: 

Fantastic. Then can you take us through at least one journey where maybe it was started a year ago, you tell us, but the gradual process, how long it took, what could we learn from of the bumps and bruises along the way, or what worked really well? 

 

Becca Buell 0:31:18.1: 

Absolutely. I would say from an end-to-end, I've referenced this AI lab concept, one that worked really well was identifying specifically what part of the process, originally we were looking at cashflow, balance sheet, and the P&L, and it was quite a bit to take on from an AI overall like analysis and picking specific use cases. So and also annual planning, quarterly forecasting, and monthly forecasting. So what we did was identify what part was the most time consuming to put together, okay the quarterly P&L was, we then layered it down to what part of the P&L is the most volatile or the biggest contributor, workforce planning. Then we identified that it wasn't an accuracy problem, or an insights problem, it was an efficiency problem that we needed to solve. So all of our discussions with planners really focused on how to drive more intelligence, and I think some people assume it's in the planning models themselves, but actually one of the more interesting use cases that we're helping the team with is actually there's an 18 month procurement project that's going on, to help improve the quality of the vendor data that the FP&A team relies on. FP&A can't wait 18 months for that data to be sound. So we actually are working with them on a mapping tool that will help them infer from journal line item entries to populate that, and then leverage the Anaplan analysts to interpret and make sure that that's actually going to be sound and reasonable. So that's one particularly interesting use case, because I think people assume it has to be something that is always machine learning, or something that is going to be in the models. Sometimes it's actually helping make the data itself from an anomaly detection and quality and cleansing in a good spot. 

 

Neil Thomas 0:33:08.3: 

Got you. Things I hear are, do we need special skills, do we need to go hire PhD data scientists, so the teams that you're working with, have you seen any change in the characteristics from maybe a traditional old implementation type of style, or is it the same folks? 

 

Becca Buell 0:33:27.2: 

I would say the folks that are within Spaulding Ridge and within the clients themselves, the makeup honestly hasn't really changed yet, I do think with time, especially as folks invest in features like CoModeler, or Agent Studio, it has to be a skillset that's a little bit more curious and business oriented, and less, probably a little bit less technical honestly from the PhD perspective, and more understanding. The folks that are going to thrive are those that can capture and interpret business process, making sure data is informing the right business rules, that AI will actually act upon. So I actually think it'll go less technical from that perspective. 

 

Neil Thomas 0:34:09.6: 

Got you, okay. We have five minutes left, we do, okay. So I could ask you too many questions, but I will if nobody else wants to ask a question. So anybody in the audience would like to ask Becca, you can ask me a question if you wish, but she's more interesting. Anybody have a question for Becca, on what they've heard so far? Or are well all very quiet, freezing to death? No, we're not, there we go. Thank you sir. 

 

Audience 0:34:35.0: 

[Inaudible 0:34:35.0] AI agent in trained on using those reference models. What if we customize the application to an extent where the AI is not able to give you the right results. Applications will probably work 60 or 70 per cent of the time, there is always some part of customization you have to do, to have it fit your business process. In that case would that AI analysis, or the AI models still work and give you the right results? 

 

Neil Thomas 0:35:06.5: 

Sure. Do you want me to answer that, do you want to answer that? You get to use the professorial thing, so you'll look intelligent if you stand up and use that. 

 

Becca Buell 0:35:15.7: 

Just so I'm understanding the question, how can you make sure that the agents are interpreting your custom data models instead of maybe something that's built out of an application, is that fair? The way we've been approaching it with organizations that are doing that, where they're building the agent on top of a bespoke model is the legacy term, is picking the most common questions that planners get stumped with, or that they're asking the COE and it's… I'm sure folks that are in parts of COE in this room, you get the same question almost every cycle, especially if FP&A team is growing and you're trying to figure out how to triage while also working on bug fixes and backlog. So what we're doing is testing those questions to make sure that it's learning and pointing to the right pages, and where it's not we're going back to where incremental data needs to be added, or it needs to be structured, so it's actually like going back to fixing and making sure your data model is streamlined. Some of that takes time, we're in early stages of testing how much iteration that takes at this point. Happy to talk more once we get a little bit further along on that. Do you have another perspective? 

 

Neil Thomas 0:36:31.1: 

Yes, and you heard about the Agent Studio product, that's what that does for you. So we use that to train the applications, if you implement an application you get 200 something questions immediately, and once you load your data in it's already smart. If it was a bespoke model, never see me, you'd get the same tool and you would go through the same process. So it's another advocation benefit, but if you're very comfortable with your data and your structures to Becca's point, I can do it, so it's that easy. So you just go in, ask the questions, check the answers, and yes there may be a little bit of upstream need to change the model a little bit, make sure it's working, change my dataset, but that's helpful, right, that's helping find things that you can improve on anyway. So yes, watch out for that coming soon. There's a demo going on somewhere probably around that today. Yes, sir. 

 

Audience 0:37:28.4: 

Hey, what was the longest implementation time for any of your client? 

 

Becca Buell: 0:37:36.6: 

The longest what time? 

 

Audience 0:37:36.1: 

Implementation time. 

 

Becca Buell 0:37:38.2: 

Oh the longest… 

 

Neil Thomas 0:37:40.6: 

Do implementations ever stop? 

 

Audience 0:37:40.9: 

For finance side, for finance applications. 

 

Becca Buell 0:37:43.7: 

That's a good question. I would say if you're looking at a crawl, walk, run approach, usually I recommend launching the longest first use case in the span of like under four to five months, so that way that you can start to get value from it. When it stems beyond that, there's probably business processes that are changing within that five month period that you need to revisit the original design, and the agile nature of Anaplan is diffused a bit. I've certainly seen implementations for EPM go much longer than that, where it's you're not doing the crawl, walk, run, where you want to implement a more programmatic every use case that will contribute to a sub model for the P&L for example, and then you want to feed in a little bit more granular project planning to contribute to the cashflow. So that could be closer to an 18 to 24 month EPM implementation, because you're really maximizing and taking every component. I don't know if I have a good theory for how long the AI EPM overlay would be, our current roadmaps that are executing against we're taking them, and the longest is usually a quarter, because we want to maximize the value before the tech, so we're learning from the technology as it continues to change. 

 

Becca Buell 0:39:03.2: 

As you guys probably know if you're keeping up with AI it's changing all the time, like Anaplan's AI features are changing, and continuing to add additional ones. So you don't want it to go much longer than a quarter is my take. 

 

Audience 0:39:15.7: 

Okay, thank you. 

 

Neil Thomas 0:39:20.4: 

Any other questions anybody? Oh, chap at the middle left here. If everybody had a microphone it'd be so much more convenient! Thank you for the question. 

 

Audience 0:39:36.9: 

HI Becca. So I'm curious if any of your customers have built their own like say Codex or Claude agents and wanted to integrate that with an existing Anaplan application. Have they come up with questions like that and asked you how to go about it using MCP server or whatever route you can propose? 

 

Becca Buell 0:40:04.0: 

Absolutely. I've seen, especially with some of these AI labs we're approaching it often from a tech agnostic lens, and seeing within not just your EPM or in this… Your Anaplan ecosystem, or if you have a hybrid EPM environment. There's also AI, like I mentioned, that could be required upstream in your ERP, or in your CRM, or even your analytics that are intertwined within Anaplan. So I've seen plenty of applications where it's a hybrid of like multi-tech, and I think it's good to trial that, to then identify what's working and what's not, where you have resistance in the organization and where you do not. Where the data needs improvement, so it's going to be of value, even if you eventually throw it away, because it will either become stale or you want to invest in the platforms, like you already have that are infusing their own AI into them. 

 

Neil Thomas 0:40:59.0: 

We have certainly architected it from a perspective of expecting that to happen, and we have seen the beginnings of that, where a customer says, 'I'm a Gemini company, I never want to use anything else.' You obviously still need to leverage the engine underneath, but the query comes through that interface. So we use a model context protocol, it's totally API driven, if you are willing to make that investment you can absolutely take advantage of that, so yes, yes is the answer. I would say even in the last, for me, last two or three months it's gradually, the interest in doing that is picking up, because of course customers want to standardize whenever they can. So it's been a deliberate design from us to try and make that as easy for them to do if they wish, as possible. 

 

Becca Buell 0:41:44.9: 

I am seeing some folks start to migrate from potentially a homegrown or leveraging for example Claude or a ChatGPT that was almost like an FAQ for planners, and now they're looking at how can they reduce the friction of a planner needing to leave their planning models, go outside, go back in, and have it all, like a migration of the prior solution into Anaplan. Yes. 

 

Neil Thomas 0:42:11.1: 

Yes, still early though, so it's going to be interesting to see, but definitely seeing that. Are we out of time, or any more questions? No, I don't see any hands, so I think we are out of time, thank you very much. Becca, thank you for spending some time with us today. 

 

Becca Buell 0:42:26.8: 

Happy to talk more around AI labs or whatever at the team. 

 

Neil Thomas 0:42:30.3: 

Becca would love to give you all an AI lab before the end of the day, is that right? Yes, there you go. 

 

Becca Buell 0:42:34.3: 

That's it. 

 

Neil Thomas 0:42:35.3: 

All right, thank you so much for spending time with us, have a great time today with us, thank you very much, thank you.