Jeremy Diamond 0:00:13.1:

Good afternoon. We're live. This is a hot mic. It's loud. Surprise. Okay. Good afternoon, everybody. This is the final but best event of the financial track. For those of you that have been with us this afternoon, my name is Jeremy Diamond. I'm the affable MC for this finance track. We've had similar themes, right? The increased pressure, the need for real-time productivity tools, the need for robust enterprise-grade tools that can't break. And going through financial transformation at the world's largest bank all of those things are true, right? This is not playtime. This is this is the real deal It's also apropos that I get to introduce the panel at JPMorgan Chase, my name being Jeremy Diamond Sounds a little bit like Jamie Diamond. We have some fun with that in in the office and there's some positive confusion. Positive for me, maybe not so much for Jamie Diamond. JPMorgan Chase's global CFO's name is also Jeremy Barnum. So you can see how sometimes the scheduling gets mixed up and we've had a lot of fun with that. So without further ado, it's my pleasure to introduce PJ, who is with Anaplan on our banking team, covers JPMorgan and some of our other strategic accounts, and Fash, who leads transformation and the COE at JPMorgan Chase. Thanks for being here.

Fash Dabesh-Khoy 0:01:45.9:

Thank you.

Paranjay Malhotra 0:01:49.7:

Thanks, Jeremy. Well, welcome. It's great to have you. I'm honored to interview somebody with such great taste in hairstyles. So I think, Fash, just to start off, do you mind telling everybody just a little bit about yourself, about your role at JPMorgan and how you contribute to being here.

Fash Dabesh-Khoy 0:02:10.4:

Yes, absolutely. So thanks for having me. It's great to be here. And I recognize this is the last session of the day before happy hour, so we'll try and make it as exciting as possible. So I'm Fash Dabesh-Khoy. I am the head of P&A transformation for the firm. And I basically run a team of data modelers, forecast solution builders, and product managers in various different groups where we aim to transform the way we do planning analysis at the firm, partnering with different lines of business. And I think we're going to go over all the different details of that. But yes, it's been a journey. I've been in this seat for about four years and been thoroughly enjoying it.

Paranjay Malhotra 0:02:54.7:

We're happy to have you. It's funny, I think you were up here about three years ago, right?

Fash Dabesh-Khoy 0:02:59.9:

That's right.

Paranjay Malhotra 0:03:01.0:

A little different part of your journey.

Fash Dabesh-Khoy 0:03:02.9:

Very early stages.

Paranjay Malhotra 0:03:03.6:

Yes, very early stages. So I think one of the first things I wanted to just talk about is transformation has been evolving pretty heavily, especially with the way the world is changing. Can you talk about what does transformation at the scale of JPMorgan look like?

Fash Dabesh-Khoy 0:03:21.4:

Sure. So I think we started the journey because I think at the time, people realized that there is going to be a need for more timely, more granular, more detailed insights. that's enabling us to make decisions. It's all about predicting our future in a way. So we just knew the direction of travel is just going to be faster, more detailed, more granular, more insights, faster decisions that we need to make. And we basically set up the transformation agenda. And the way we structure the transformation, and it's really in a simple term, it's actually three pillars, the way we think about transformation. One of the key things that we actually looked at was, number one, which is data. So you look at many P&A functions around the world in different companies, the planners, the people doing the analysis, they actually spend the vast majority of their time just trying to find the data, source the data, reconcile it, join it, transform it, to get it in a shape that actually enables them to do their planning and analysis. And our thought was, well, if we can make a dent in that, we can actually have a pretty big impact. The second pillar that we looked at was, okay, so the data that planners and people doing analysis need is there. Let's say we get to a point where the data is there. They're going to need some tools, some applications, some modern platforms in order to do forecasting. Hopefully getting them out of their Excel files in some shape or form. We know people in finance love Excel, so we have some tactics to try to please them in some ways.

Fash Dabesh-Khoy 0:05:02.2:

And then the third thing that we looked at was, and we're going to talk about how sort of Anaplan plays a role in that forecasting platform for us. And then the third part was, okay, so you've got input data, you've got models, you've got platforms and tools that you're doing forecasting, there's an output. Okay, when you get that output, how do you then analyze it? How do you actually enable the finance organization to analyze it? And what's interesting is when you look at the finance organization, there's a massive range of people with different skills, right? So you get the people who are writing Python and SQL and machine learning models and doing some seriously complex data science work. And then you've got the ones who are like, yes, I like pivot tables in Excel. And you get everything in between. So our consumption strategy was really around how do we enable everybody, meet them where they are in terms of their consumption of products. And that was the third pillar of our transformation effort that we have been on. The biggest one, obviously, most important is data. That's actually been truly the foundation of the transformation journey for us.

Paranjay Malhotra 0:06:12.7:

That makes a lot of sense. In practice, what does getting the data right actually mean? Obviously it's having access to the right data, but what steps are you taking to make sure that that data is the right data, standard data, governed, and what's the discipline you've been instilling across your organization?

Fash Dabesh-Khoy 0:06:31.6:

Oh yes, I mean it's been huge. It's been a massive journey of discovery for us, so I'll tell you, we just knew we had to get our data right, but we had to figure out, okay, what does that mean? So we have a data platform that we have deployed. And on that, what we do is what we build is data products. So we treat each data set that we use as essentially a product. It has a lifecycle. It has requirements. It has sources of data. We've been building automated pipelines to bring the data in. We build controls in terms of data quality, timeliness, refresh cycles. Reference data has been a huge pain for us, right? Cleaning that up, hierarchies, cycles, all sorts of various different ref data that was kind of all over the place, and everybody was taking a copy and doing it themselves in their Excel files. We actually stood up a data modeling team, a team of data architects who spent all day long designing these data products, metadata tagging them, cataloging them, thinking about data quality controls, validations, reconciliations, and so on. The idea there is that when you have a data product that we will call gold standard, anybody using it is a trusted data source that they know has been validated, is accurate, and is clean, and it has the dimensions that they require, right? But what you see with these datasets is that they actually go through a journey like any product.

Fash Dabesh-Khoy 0:07:59.3:

You may release an MVP version of the dataset, your users start using them, and you get an onslaught of requirements. 'Oh, I also need this attribute. I need that. But I need it weekly, and it's currently monthly. I need it refreshed more frequently.' So we just go through that process and get these data sets in a clean state as much as possible. And one of the key things that we have evolved to on the data journey is, and there is no fireside chat without the mention of AI, but obviously the way we think about it, working with our own chief data officer of the firm is the data needs to be AI ready. What does that mean? That means that data is modeled, constructed, tagged in a way that AI models can look at it and understand. There is context, there is detail. Without that, you're going to be hampered a little bit on how much you can get through in terms of data. And I know we're going to touch on that a little bit as well.

Paranjay Malhotra 0:08:53.8:

Yes, I mean, I think the access to data is one thing, but if the data is not right, it really doesn't make sense for anybody, right? So, once the data is right, you mentioned consumption as one of your key pillars, right? So even before we get into doing anything with that data, can you talk a little bit about your strategy on consuming data and how Anaplan has helped drive that consumption strategy as well?

Fash Dabesh-Khoy 0:09:20.9:

Absolutely. So obviously when you're building Anaplan models, one of the key things that you need is input data, starting data. And for P&A, it's typically actuals, right? It's actuals, hierarchies, and humans providing inputs and assumptions and so on. So our strategy from day one was, okay, well, we've got to make the data easily available for these Anaplan models, for the modelers who are building these sort of various implementations to be able to ingest that data. So we spent some time actually building pipelines between Anaplan and our data platform. We want to minimize people having to manually upload data into Anaplan, and I've got my own Anaplan team who've built some very innovative solutions how we bring data into Anaplan, how we reconcile and validate to make sure the data arrived fully and complete and so on. Within Anaplan, when the planners are actually in there, Anaplan is one of our areas where you can consume data and actually look at it. The point there being you've got the planners in there, you've got people who are doing inputs, you've got people who are looking at consolidated views and doing top size and adjustments and so that is a little bit of a workflow. So our strategy has been when your data is inside Anaplan and you are planning, Anaplan is your lens into that data.

Fash Dabesh-Khoy 0:10:45.0:

And we've also built pipelines where when you are done, you hit a button and the data is published back out to our data platform for broader consumption for any of our planners or downstream submissions and so on. So again, we've met the users where they are. The planners can use the screens and views inside Anaplan. Downstream consumers, consolidators, the CFO of the firm who wants to review and so on, we've got a suite of various consumption patterns on top of, in addition to Anaplan that people use.

Paranjay Malhotra 0:11:17.3:

Yes. And you mentioned meeting the users where they are. Not getting into the change management aspect of this, but one of the key things that you guys are doing, which is very, very cool, is the use of Anaplan Excel to help drive meeting the users where they are because of the sometimes lack of wanting to get out of Excel. Can you talk about that a little bit on that consumption stream?

Fash Dabesh-Khoy 0:11:39.8:

Yes, I mean, look, since day one, like, genuinely, we love Anaplan Excel. It has been a game changer for us. And it's interesting that the strategy around Anaplan Excel was mostly driven directly by the CFO of the firm, the CFO of JPMorgan, who said, look, over the years, I have not seen another software application that you could put in front of a finance person and they would say, yes, this is a better interface, easier to use, faster than Excel, right? And so how could we adapt our strategy to enable users to still use Excel as an interface but we don't do some of those bad habits like the data living in the Excel files and getting emailed around the company, right? And that's when we actually came across Anaplan Excel at the time. I mean, we have done a tremendous amount of work with your product team in terms of enhancing the product, building inputs and adjustments and all sorts of performance, and the product's come a long way, and we're very happy with how it's evolved. But the way we treat it is the data is in our data platform. We actually have also put in where the data set is massive. So you look at data sets like allocations and so on. You've got billions and billions of records, multiple cycles, multiple dates and history and forecasts and so on. We actually consume those through a middle layer, which is typically an OLAP solution for us. So multidimensional cubes that actually does that pre-aggregation and pre-calculation.

Fash Dabesh-Khoy 0:13:17.0:

Then we actually connect Anaplan Excel to the OLAP solution, and the user can do retrieves, pull the data, view pivots, style it how they want, and analysis, and then they can save the template. The good thing there is that data doesn't get copied into the Excel file itself, it still remains on the cloud. So next time you open up your Excel file, you hit refresh, it retrieves the data back again. We really like that because it enables our planners to still use Excel as the interface or follow our strategy, which is around don't download data. Don't email data around. The data wrangling, the data transformation should all be done at the data layer in itself. And we're also leveraging Anaplan Excel now more and more as a front end to the actual Anaplan application as well. Some of our users and look finance people, they're wedded to their ways. There's a big change management play there and if we can meet them where they are and still instill some of these best practices in them through the use of an Excel, we're all for it.

Paranjay Malhotra 0:14:23.1:

That's great because. at the end of the day, the data could live in different spots. But you want to continue to have that governance and discipline that you've structured the data in, and then still give the users the flexibility to use that data however they see fit.

Fash Dabesh-Khoy 0:14:39.1:

Absolutely. And it's key for us. If you think about it, right, Anaplan Excel is not just wedded to the Anaplan backend, right? It can connect to many different backends. And a company our size, your data, some of it could be in one data platform and so in other data platforms, you may want to connect it to a cube. You know, there are various different sort of backends we want to connect it to. And the fact that you can query data from multiple places at once and sort of bring it together is actually very powerful.

Paranjay Malhotra 0:15:07.1:

Yes, that's awesome. So now once we have the consumption strategy in place, now let's maybe talk a little bit about planning. So maybe give, if you can, just an overview of how you're using Anaplan at the scale that JPMorgan has.

Fash Dabesh-Khoy 0:15:23.4:

Yes, I mean, look, JPMorgan, a lot of times we see ourselves as sort of a one-of-one in terms of size complexity, and some of it's definitely caused by us. But also, it's a very large company, very complex, very large set of products, and each line of business is very unique in their own way. 330,000 employees alone and many, many thousands of cost centers and various alternate hierarchies and dimensions and so on. So what we have learned over the years is that we have to be very thoughtful around how we identify use cases of planning solutions and models that we actually bring onto the Anaplan platform. Some of the key things that we focus on is the size of the data, the complexity of the calculations, and the number of concurrent users that are going to be interacting with it. And if you get those three right, you can actually churn out Anaplan models over and over. So some of our most complex implementations is really headcount and expense related. You would think it's just headcount. It can't be that complicated, but it's actually very complicated and very diverse in the way people plan. When you think about attrition and vacancy and historical averages and macroeconomic variables and various inputs, overlays, then you've got expense in terms of comp expense, non-comp expense, benefits, all these different things that come into the picture.

Fash Dabesh-Khoy 0:16:55.3:

It can become very large and complex very quickly. But I think it's been a great journey and Anaplan yourselves, your architects have been great partners in really helping us navigate through that. We've also hired some very strong talent into the team who are really forming these models in the most optimal way, performance tuning them and so on. And we've got, I would say a dozen or so models that are live now and with more in the pipeline. And we've also got some of the lines of business have actually hired their own Anaplan builders, and they're a lot more self-sufficient. They can build and deploy their own models while complying with our controls and governance and entitlements and all of that.

Paranjay Malhotra 0:17:42.8:

In terms of scale of JPMorgan, just from an expenses alone, that's 105 billion in expense, 19.8 billion in technology expense. The scale is massive and obviously that's a target state that you want to be more effective in managing those expenses. And so all of this kind of plays into doing that in a more efficient way, correct?

Fash Dabesh-Khoy 0:18:06.4:

That's right, that's right. And when you implement these Anaplan models, it enables us to drive some degree of consistency, right? So if you look at it from a firm-wide standpoint where you're consolidating data, what our CFO calls, as the data makes its way to the center and gets aggregated and goes up the tree, there's a degree of what we call data destruction that happens. Some of those attributes, some of those assumptions and inputs and so on get lost. So when you then ask a question at the top, you go through this phone a friend kind of construct to try and get to that answer. And that's where using planning tools where we can actually see the inputs, the assumptions, the user actions, the clicks, like how did they get to the number that you've now got to the center? Tools such as Anaplan and any other solution that we think about in terms of forecasting, if they can capture that level of granularity and detail, it can really help us self-serve a lot and be able to more easily challenge some of these forecasts that we get, which human nature can be a little bit too optimistic sometimes and so on. So it goes back to the data strategy around dimensions, granularity, and so on, which tools such as Anaplan can produce out of the box for us.

Paranjay Malhotra 0:19:22.5:

And that's the granularity is really the bottoms-up. And you mentioned challenge. So there's also the top down process that tries to deviate, not deviate, but realistically challenge the results of the bottoms up forecast versus the top down. How is that coming together?

Fash Dabesh-Khoy 0:19:37.9:

Yes, and that's been something we've done over the past couple of years where you've got the bottom-up planning that happens. The planners are on the ground sort of looking at their head count, and usually they look at their sort of planning world and they say like, oh, this is my team. I've got five people coming in, five people going out, and so on. And then think of it like a tree, and it sort of aggregates up. And when it gets to the top, you go, 'Wow, maybe, are we really going to grow by that much, or are we going to shrink by that much?' So some of those things don't make sense, and then there's top size and all of that. So one of the things that we have been building is actually Anaplan models. We call them our Challenger models, which are using much more model-based forecasting. So what we do on our data platform, we would use machine learning models to predict future joint and lever rates, those types of inputs into that and various other macroeconomic unemployment rate forecasts and various other things. And then we use those, say, percentages to be inputs into models that are inside Anaplan, which kind of auto-generate a headcount forecast based on our headcount actuals, historical trends, known future events, and it actually produces a forecast. And then we can look at ranges. And I think the way we think about it now is like, that is like a more realistic forecast that we can stand behind, and it actually enables us to have much richer conversations with our finance organization to challenge where their headcount forecast is outside these brackets.

Fash Dabesh-Khoy 0:21:12.5:

Because sometimes it's valid. There are business things that the model doesn't know about. There are events that the model doesn't know about. But that's where the top-down meets the bottom-up. And it's enabling us to have those much more directed, specific conversations versus before it was much more sort of holistic and, 'How did you get to this?' 'Oh, well, the phone a friend' type stuff. This is all encapsulated under the transformation agenda that we're running and just driving that enhanced transparency and just the elevated level of conversation that occurs.

Paranjay Malhotra 0:21:49.8:

Yes, and I think because of these rollouts, you're really entering what I would call phase two of your journey, where it's moving quicker at larger scales and creating better decision insights. And I think, so we know everybody here has been going through all these sessions with the investments that Anaplan has made, half a billion dollars over the last few years to make our product better and roll out different capabilities. How do you envision Anaplan's innovation, the things that you have today, the things that are coming out, really helping contribute to the success of the program?

Fash Dabesh-Khoy 0:22:23.3:

For sure. Now, firstly, I see Adam's in the front row. So Adam and the team have been great partners. Hey. Good to see you. And I think definitely over the past two, three years, we've seen a huge acceleration in terms of feature build-outs and just the evolution of the product as a whole. And at the same time, I find Anaplan team very receptive to our feedback and requirements and things that, as any product person, customer says, 'I need this' and you go, 'Oh, that's actually good for my product. Yes, I'm going to go add these features to it.' So we've seen quick turnaround on these events that have helped us, whether it be some of the less exciting stuff like securities and keys and that type of stuff versus, very cool UX front-end type features that they've developed, so that's been great. Secondly, we're very excited about Polaris. We are embarking on building our first Polaris implementations because we're finding that some of our models and use cases deal with pretty sparse data, and we think that we can get better performance, more efficient sizing and more concurrency out of Polaris, so we'll see how that goes. We've got some use cases that we're thinking about with the team that we're going to implement this year. And then the last thing is really for me, what you're doing on the AI space is actually great because what it gives us, at least on the co-modeller element, is time to value. So some people say patience is not one of my strong traits, so we want to try and get models out as quickly as possible.

Fash Dabesh-Khoy 0:24:06.6:

And if we can use AI to help our modelers and builders build faster, and as you know, JPMorgan is leaning in pretty heavily on the AI front, especially within our tech space, especially within our software engineering environment. We now have these AI models already integrated into our IDEs, where our software engineers are building software, automated unit tests, all of those not so exciting things that you have to do, as long as it helps us accelerate the build and improve the quality and the feature rich implementations that we can get out of it, we're all for it, so we're excited to see how the AI features evolve as we go forward.

Paranjay Malhotra 0:24:50.3:

Yes, I think recently Jamie Diamond was talking about how AI has rolled out across the firm at great scale actually. And I think we talked about it recently where one of your target goals with AI is not necessarily just full displacement type of thing, but it's removing the no joy work out of a day-to-day process. Could you talk a little bit about how you're thinking about that transformation versus just pure automation type of thing?

Fash Dabesh-Khoy 0:25:18.3:

Yes, absolutely. And I think our focus has really been around productivity, right? How do we actually improve the productivity of our folks? And the no joy work, that's actually a term coined by the CEO of our asset wealth management business, where she's like, look, there are some things that we just have to do that we don't enjoy. If you think about our controllers, applying the same adjustments and updates month after month after month. And there are situations where you're making a decision between A and B based on a certain series of conditions and so on. Could you deploy agents that do that? Could you kind of drive some level of automation around rule-based, or whether it's agentic, whether it's just classic data science or machine learning? It could be anything. Where can we do those things? That's productivity. What we mean by that is like, we want to free up our planners to do more analysis and more thinking and more decision making versus inputs and adjustments and updates and so on, which is why as we get more and more Anaplan models out there, what I'm spending time with the team doing is really analyzing the behavior of the planners, the number of changes they're making, what are they changing, how material are these changes, and so on, and really learning from that behavior to see how we can be more efficient. In terms of the top-down message from our CEO and the general theme, look, when we started to talk about this event, I feel like so much has changed with certain AI models that have really turned the table a little bit.

Fash Dabesh-Khoy 0:26:50.5:

And I think it's such a fast-paced, fast-moving space. And I think we are investing, we are leaning in heavily, and we are actively continuing to look at use cases. And you guys heard in the company update, you may try 20 use cases and five of them will go great and 15 won't, and that's okay. So we want to make sure these use cases are real. and they are adding value and that we continue to learn and evolve going forward.

Paranjay Malhotra 0:27:19.7:

Yes, yes. It's probably a question everybody's trying to answer today. What are you thinking about how do you intend to measure value? Because it's not really an arbitrary number, right? There's many things that you look at from a productivity value gain. How do you measure actually things that you're doing in Anaplan as well as the future of AI rollout?

Fash Dabesh-Khoy 0:27:42.9:

Yes, I mean, examples of that would be, for example, where let's say you have an Anaplan model that forecasts a particular sort of financial output, let's say your expenses or your revenue and so on. Wouldn't it be great that, imagine, your actuals have just landed, an agent has validated the data, the data has already been loaded, validated, a model has already generated a forecast for you, and you get a note saying, your forecast is ready for your review, right? So you just go in, you can do your top size, you can make your decisions, I'm good with this, go. Versus, ah, well, let me go get my actuals. Let me load it in. Let me reconcile it. Now let me start putting a bunch of assumptions and things in and do the forecast and then manually export it. Like, that journey we just said, the person is looking at a pre-generated forecast. They may be able to tweak some of the inputs and rerun different scenarios in rapid succession, and then make the decision on which one they think is the right one they want to publish and just go from there, and really then focus more time on commentary and so on. Then imagine if it goes even further. It does some various analysis to your actuals. It does some commentary about why it varies to your actuals and so on. People don't enjoy writing commentary and so on. I get it's a good important control point, but these are all the no joy work that I think we can actually have AI and automation around to make things as pleasurable for our financial planners as much as possible.

Paranjay Malhotra 0:29:18.1:

And then lastly on this front is, this will change the way your leaders consume information. The ultimate goal of the transformation is to be more effective with the decisions you make. How do you see that coming together with your leadership actually changing their consumption strategy as well as their decision-making strategy?

Fash Dabesh-Khoy 0:29:39.6:

That's right. You see, this might resonate for a lot of folks in the room, but historically, the patterns that we saw was what I would call fixed insights, which are like the dashboards of the world. So somebody has cache some data, built a dashboard. Sure, you can do filters, you can do some drilling, you can go to page to page, but you're limited to the realms of that dashboard. And what would happen is, you're in a meeting, somebody asks a question, just happens to be something that's not available in your dashboard or your pre-canned report, and you have to go and spend two, three days, come back. And I'm not saying that's the case every time, but these are examples. And the consumption strategy of our leaders is now saying, 'Hey, look, I want to be able to navigate the data myself. I want to be able to drill and ask questions.' So that's where we have situations where we have solutions which are a little bit of low code, no code sort of stitching of data together. And also solutions such as Anaplan that can actually bring data in, in terms of like your model that you have created a forecast where people can navigate, apply filters and so on. It's a great consumption tool also. We've got other stuff like conversational analytics, which people can prompt and ask questions on top of our data platform. But the point there is that you're only limited by your entitlements because the data is on the platform and you can actually navigate and join data and drill and essentially derive your own question.

Fash Dabesh-Khoy 0:31:13.1:

Now, but that raises one other challenge for us. It's not all roses and flowers. It's more like how do you know that you've done it right? How do you know that you've arrived at the right answer? And that's where our message everybody that we have is make sure you're validating your query. Make sure you are double-checking it. Make sure you are consulting with the data expert to make sure the insight that you have arrived at is accurate.

Paranjay Malhotra 0:31:39.4:

We talked about this journey. You're getting there, and it's evolved pretty drastically. One of your favorite topics seems to be, well, maybe not favorite topic, change management, and all this. That's probably one of the most difficult things, especially at the scale you're doing it at. Advice on change management or how you're thinking about change management?

Fash Dabesh-Khoy 0:32:03.6:

I alluded to this earlier, but like, finance folks are really wedded to their ways of doing stuff, right? Their tools, their spreadsheets, their templates, it's reality, right? And the joke that I use internally and there's some truth to it, it's like everybody wants change as long as everything stays the same. And that's where we have to like really spend a lot of time. We sometimes have to use the carrot and stick approach where we can basically say, okay, look, here's a much better way to do this. Here's the data, la, la, la, la. And some people adopt and some people are very lean in and adopt. And the others will say, we're literally removing your access to this other way you get the data, and this is now the only way you can do this, right? So we prefer the former, of course, but sometimes you're trying to turn off legacy tools, legacy applications, you have to drive that. But what I would say, though, is, and I'm sure many of you here have had top-down managerial support to be Anaplan customers and to make the investment and go on the journey. Having that top-down managerial support, genuine direct sponsorship from the CFO of the firm, whether it's in his town halls and sort of various talks and emails and so on, has been truly instrumental to us. And when people see that the CFO is in there clicking around and doing things, it definitely inspires them. And I think lead by example has been a big thing for our senior leaders, but that top-down thing has been super instrumental.

Paranjay Malhotra 0:33:43.6:

That's awesome. From an advice perspective, I would love for you to give advice given that it's three years back on stage now, a much different part. For people beginning their journey, people evolving, what advice would you give to everybody here, especially as they see all the innovations and things coming?

Fash Dabesh-Khoy 0:34:00.2:

I was walking around earlier, spoke to a few people, very encouraged with common problems, common pain points that we were trying to solve. I think now having been on a journey for some time, my suggestion would be don't lose sight of the data, right? It's the core foundation of all of your planning that you're going to be building in tools such as Anaplan and so on. It's grunt work. It's what we call with my boss, eating your vegetables, but then you do get some dessert at the end, which is the AI stuff. But you've got to get your data right, that's a big thing. Patience, stick to the journey. You may find challenges and hiccups along the way. We've had to rebuild and redo certain things and it's been a massive learning experience for us and it's been enjoyable for sure. But patience is the key thing. And then just continuing to push on adoption. You've got to continue to push, encourage, and motivate, and market, and try and get the word out. Try and create buzz. We run hackathons. We run massive training stuff. We do videos. And we try to create an excitement to bring people along. One of the tactics is doing some kind of FOMO. Well, they're using it and they're really happy, so why aren't you? So it's a little bit of that too. So you have to, it's multi-dimensional in the approach that you take, but data, patience, and just perseverance in terms of driving that change management into habits.

Paranjay Malhotra 0:35:35.1:

Yes, and we've been partnering through this together. So I think we have a very good partnership to make sure you're getting the success that you're hoping to get. I guess I'll put myself on the spot and the team here, but where's areas in our partnership that we could improve on to help drive even faster time to value, even more value from a decision-making perspective? Just our relationship in general. I won't be offended.

Fash Dabesh-Khoy 0:36:06.0:

Adam's listening now. No, honestly, you guys have been incredible partners. It really goes without saying. I think we're able to just pick up the phone and call you guys when there's an issue, when we need something, and the response is just fantastic. And I think for us at JPMorgan, we genuinely feel that the Anaplan management team, the product teams, they genuinely care, really care about our success and have been great partners. We meet regularly. We have a voice in terms of the sort of evolution of the product. We're very encouraged by the investments that you're making in there and so on. I think for us, in terms of things we could continue to do, I would say, is one, we've seen a lot of discipline and focus in terms of investing in the current platform, whether it be user experience, whether it be core foundational components, whether it be security and so on. Fantastic. We want to continue to see that happen. Because we do have live use cases, live models that are running, so that's great. The other thing I'm super encouraged about and excited about, we've been impressed with some of the demos from the AI features of our co-modeler and the AI for the, was it Forecaster?

Paranjay Malhotra 0:37:20.6:

Forecaster, yes.

Fash Dabesh-Khoy 0:37:21.3:

There we go. See, I remembered. Forecaster and so on, so we want to see more of that. And also the other thing that I think it will evolve and I'm looking forward to our next conversation with Adam and the team is like, well, agentic AI, what role do that do they play as they interact with planning solutions and so on? Like, how could we automate some of those operational processes around preparing for forecasts and doing the forecasting, using agents, how can they communicate with Anaplan? These are things that are playing on our mind, but I'm excited on the journey that we're on to see where we can head.

Paranjay Malhotra 0:38:02.6:

Yes, I think they're on Adam's mind, too.

Fash Dabesh-Khoy 0:38:05.3:

Adam's giving me the thumbs up. Maybe I asked the right question.

Paranjay Malhotra 0:38:08.3:

Thanks, Fash. I think we actually have a few minutes left. If anybody in the crowd wants to put Fash on the spot and ask a question, have a mic open. I know it's the end of the day, so it's...

Audience 0:38:34.7:

So when you talk about the integrity of the data and kind of pushing improvement in reference data, et cetera, finance often is the person who identifies these data issues. And it took a while for us to be able to really push those data issues upstream, even just having a loop feedback loop where, hey, I saw this didn't look right. Who do I ask? Because we're just feeding the data from everywhere. How have you kind of solved that loopback stream? And even now that we have a process for that, even just getting our finance people to have the willingness to want to push that accountability back where it belongs instead of just, to your point, coming up with their own little table or their own little mapping or their own little something that fixes that problem. How did you do that journey to get really comfortable that that data coming to you is correct or have an ability to fix that back up?

Fash Dabesh-Khoy 0:39:22.9:

Oh, great question. So we do it in we've done it in two ways, right? So one is there are sometimes data sets that we're ingesting that comes with gaps or missing attributes and so on. And to your point, sometimes it is because of upstream, maybe the transactional system that captures the data, maybe the tool or the process that manages the data has gaps in it so the data just is produced which is in a way you cannot trust. So part of that is really just going upstream and getting senior sponsorship to prioritize the fixing of those things. But they can take some time, right? They do take some time. So the flip side of that is like, well, what can we do in the meantime? So some of the tactics we have used is like, well, not all the data is bad. Some of the attributes and things are pretty complete. Why don't we at least bring those in? Let's start using those with a high degree of trustability onto the data. Two, could we possibly derive things? Could we actually use some reference data to auto-calculate and auto-derive? We call it the term stitching. Could I actually at least bring some other data in here and stitch it together to actually give me the view that I want? And obviously reference data is a huge point that you mentioned also. We actually identified owners for each one of these data sets and we made them accountable. Like your performance is related to actually fixing this data.

Fash Dabesh-Khoy 0:40:57.5:

And doing everything fully strategic, solving the data in the core systems upstream, again, does take a long time. But we've actually implemented processes where they could input the data, even if it's through Excel, whatever it is, to actually make it complete, and automated reconciliations, automated data quality controls. But it's not a silver bullet in terms of fixing those. So we have some data sets which we have resolved, but still conceded that you're still going to get a handful of breaks and things, but at least we've got the exception reports that tell us what's not complete with the data, so at least when we provide the data, we also give them the exception report and say, well, this is five records that are still being validated, but these two million are okay, right? So it's the top-down messaging you have to do with the upstream systems, which are mixed levels of success so far, but it's a journey, the patience thing was in there, so you really have to continue to push.

Paranjay Malhotra 0:42:10.5:

Anybody else? Question?

Fash Dabesh-Khoy 0:42:14.5:

I think everyone's ready for a drink.

Paranjay Malhotra 0:42:15.4:

Yes. Going once, going twice. Overall, just thank you so much for your time again. Hopefully we'll be back up in another three years from now, and your journey will be way further with all $105 billion in expense. 

Fash Dabesh-Khoy 0:42:35.9:

Yes, we're excited. It's been a great journey, and thank you for everybody who came out today. Hopefully this was useful, and thank you for the continued partnership.

Paranjay Malhotra 0:42:45.2:

You got it. Awesome. Thank you.