Viji Doraiswamy 0:00:02.9:
Good afternoon, everyone. It’s great to see a full house here. Also, I’m so excited to start off with an energizing topic like AI and planning, which is going to make it challenging for all of you to give in to that afternoon slump, so I hope so. Nobody is going to get a post-lunch siesta here. We have something really exciting to talk about, for sure. Now, before I indulge in the topic itself, did you know that AI was announced as the most notable word of 2023 by Collins Dictionary? Any surprises there? I don’t think so, right, because especially with the whole ChatGPT revolution that spurred in, towards late 2022, every company, every software company and every other company in the world wanted to jump on this AI bandwagon. But then we all know that AI is much more than just a buzzword. We know that it has significant power to transform business processes for enterprises, and especially in the world of connected and business planning. Luckily for companies like Anaplan, we have been on the AI journey long before it became a buzzword, for several years now. That allows us the opportunity to build on our foundations and investments and to tap into these latest technologies to deliver value, and we have seen this also being tapped and used by several of our customers that we are excited to share about today.

Viji Doraiswamy 0:01:30.9:
Before Lesley and I jump into talking of giving you a sneak preview into our vision and our road map, what’s coming in the future as well as, and how to make you, as enterprise planning leaders, the leader of tapping into AI to deliver value to your organizations. Before we get there, I’d like to know from a show of hands here, how many of you are actively considering or pursuing AI pilots or use cases in your business organizations? It doesn’t have to be on Anaplan, but if there are things that you’re testing out or pilots that you’re running, how many of you are already doing that? About a third, I would say then. Now, would anybody like to share what kind of use cases that you’re looking into at this point, like, that is related to business planning with a sales planning, supply chain planning, finance planning, forecasting, any of those? Okay, it’s a small crowd here so you can, like, at least in what area are you doing, or are you looking at pilots at this point? There are one or two, yes.

Audience 0:02:37.7:
[Unclear words 0:02:37.9] statistical forecasting.

Viji Doraiswamy 0:02:41.8:
Statistical forecasting use cases, typical.

Audience 0:02:45.8:
[Unclear words 0:02:45.8] results outside of the current standards [unclear words 0:02:53.6].

Viji Doraiswamy 0:02:54.7:
Excellent, yes, thank you. Anyone else? Yes?

Audience 0:02:58.9:0:
Specifically for [unclear words 0:02:59.0] something where you can ask – [receives microphone] oh, thank you – somewhere you can ask and retrieve data, particularly for your business partners who, looking at [?everything else 0:03:08.9] as part of their finance concierge, to provide data on demand to support presentations and be able to do that using natural language.

Viji Doraiswamy 0:03:16.1:
Excellent, yes. I think that’s a standard typical use case that we see as well. Anybody else like to share? Okay, well, we’ll discuss a few use cases here and what companies are doing, and how you can start taking advantage of that. I’d like to start out by putting out Anaplan’s AI vision. Now, if you look at Anaplan’s AI vision and journey, it’s centered around three big goals. The first one, obviously, is to make sure accessing AI enabled data and insights in the hands of everyone, starting with C-level leaders to our directors or heads of business planning, to casual users, and especially with generative AI coming into the picture, this makes it easier for us. You saw Adam talk a lot about Natural Language Querying this morning, so using that, how can we enable people to query? It could be as simple as, depending on the persona, asking for what was the revenue numbers last year, to all the way to, okay, what is my forecast for this year, and how can I enable it better? How can I tweak that better to achieve my goals all the way, so it’s enabling AI access for everybody.

Viji Doraiswamy 0:04:26.9:
Secondly, deeper insights with AI. Now, AI has shown us as even we are doing all of these pilots, that it has the ability to connect data models and the platforms, be in ways that we have not been able to even crack it with human intelligence today. It can in the background, really go beyond systems and connect the data, identify those patterns and give us more deeper insights, and we talked a little bit about how this is happening today. Then finally, I would be remiss if we don’t talk about how Anaplan taps into large language models and use it for, to increase the efficiency when it comes to model building and model generation. Now, in my role, I have the opportunity to talk with a lot of our analysts, like across Gartner, Forrester, IDC, etc., and this is the number one use case that they come to. If they see a big, single, big differentiator for Anaplan, it would be in that model as using AI to assist in the model building phase, because we see how much time it takes to build and generate the complex models.

Viji Doraiswamy 0:05:31.6:
If we are able to free up the time of our model builders by cutting down that time drastically, based on experience of building and tapping into models from across various companies within an industry, or even outside of an industry, if we are able to use that, that’s the true value that we could deliver for our customers as well. These are the three big objectives for us with our AI vision. As I mentioned before, we are not new in this AI bandwagon. We started this journey back in 2018 when we launched Optimizer, which was basically linear programming optimization algorithms, but what it was able to do was, we were able to set, define a goal, whether it’s cost reduction, whether it’s profit maximization, whether it’s increasing revenue. But you’re also able to set a few constraints around it, and it’s able to define and come out with more optimized models for you to choose. We’ve seen this used in things like quota planning and sales planning etc., where you’re able to set constraints, maximum number of accounts per sales rep or minimum number of accounts. If it’s a major account, then it’s just one account per rep, etc.

Viji Doraiswamy 0:06:43.1:
Whatever constraint it may be, you’re able to set that and use – you’re able to set goals and able to optimize that for your business benefits. Then in 2020, we came out with Predictive Insights through an acquisition, but basically combining your historic revenue and your CRM data with third party market signals, and able to give you the insight into which accounts you need to target and engage with. We have seen this work for a few of our customers who’ve also said that historic data sometimes only tells you so much, but with… When you’re able to combine it and you’re able to identify which are the accounts which have the higher propensity to buy, and that’s your true benefit there. In 2022, we launched PlanIQ. Now, Lesley will talk a little bit more into this, and how he’s seen this work with customers as well, but we have seen this. This was a forecasting tool, basically tapping into machine learning and using machine learning algorithms to improve the forecast and predictability accuracy, and we’ve seen this work over and over again as well in several industries.

Viji Doraiswamy 0:07:51.5:
Now, based on this, where are we going after having launched these three products? You saw Adam talk in the keynote about generative AI, and using Natural Language Querying that is going to be built into several of the applications. In a second here, I’ll show you how this could work. We have some screenshots of the product that I can take you through and what this could mean for us across different applications. Also, next year, we are looking at launching an early warning system, which is basically combining, again, looking at historic, millions of data points historically, but also combining it with external sources of data that allows you to identify anomalies, deviations. And when that happens, not just giving you a warning, but also coming up with recommendations of proactive actions that you should take in order to be able to correct it. This has huge implications both in the supply chain, in the sales, as well as in finance planning space, as you can see.

Viji Doraiswamy 0:08:51.0:
Then finally, as we talked about, the model builder co-pilot. We have been testing this out and we’re already seeing some of the model generation using generating hierarchies or setting up hierarchies and libraries, the time coming down. What used to take a day-and-a-half now takes just a couple of hours, which is going to be huge when it comes to efficiency building. Now, we talked about generative AI. Now, these are some examples of what you could do, so based on the persona that, who’s using this application, for example, it could come up with prompts that is more relevant to that persona. If it’s a sales planning person, what does he or she most care about? Is it the top five revenue generating regions, or the top five opportunities with more pipeline, whatever that may be, you’re able to also query, use your Natural Language Querying capabilities, but also make use of the prompts that we have here as well. Not only that, you could also then go one step further and ask it to combine different sources of data and come out with generative analytics.

Viji Doraiswamy 0:09:57.5:
It could be, like, you’re on your comparison, or can we look at what worked last year versus which are the revenue, top revenue generating regions of this year, what could that mean, etc. So, this does make use of generative analytics in the background, crawls through various systems and pulls that data, and it actually presents to you in the format that you wish to see it. It gives you various options to pull depending on the personas, depending on the users, on who you talk to, and be able to see that. More importantly, it has the ability to trigger Workflows, so it can send it to you for approvals, which you’re able to use on your mobile device and approve that, so this is what we mean by enabling AI for all, and giving frictionless access to AI driven insights. We’ll see more about it in the second half of the year, some of these capabilities come out from Anaplan. Now, I’d like to give you a sneak peek, or this is more of a snapshot rather, of what benefits our customers have experienced.

Viji Doraiswamy 0:10:56.8:
Now, this may seem like, okay, siloed benefits that they are experiencing in their area, but these have huge impacts if you see in the background. In fact, one of our customers told us how a 1.4 per cent increase in their predictability accuracy actually led to savings of about $3 million to $4 million in the year, and not just monetary benefits. Depending on the industry, and depending on the use case, this could mean risk mitigation. In the case of a hospital system, it’s actually predicting how many beds or how many employees do you need, so it is almost like life-saving in some cases. It could mean improving your customer satisfaction and customer service levels. It could mean bringing an innovation faster to market so you can beat the competition, so depending on the use case and the industry, the benefits are varied, but this is what we’re looking at from an AI driven perspective today. With that, I’d like to turn it over to Lesley to talk about, take us through the road map as well as some of the best practices and strategies implementation.

Lesley Vincent 0:11:56.8:
Thank you Viji. This morning you all saw Adam, our chief product officer, talk about our road map, and Viji just shared our AI vision. I think one of the most interesting topics is how do we implement this in our organizations, outside of the nice technology, the strategy and the vision? How do we implement AI driven planning? I’ll start off with our road map from a high-level perspective. Currently we have explainability built into PlanIQ. What this means is when you have a prediction in the past, your AI would be like a black box. You had no way of telling why did a specific prediction or focus improved. With PlanIQ, these two algorithms, MVLR and Prophet, you will be able in a visual way to tell your vice president, C-level people, hey, this specific holiday, this specific piece of data impacted your predictions, right? That’s obviously incredibly powerful. In the near future, as Adam mentioned this morning, we’re looking at combining conversational generative AI where you could ask Anaplan a question, even on your mobile phone, get insights back, but then also be able to trigger actions using Workflow, so you get the insight and right away you’re able to trigger actions via Workflow.

Lesley Vincent 0:13:42.0:
Then from a future perspective, we’re looking at benchmarking, so how are you doing compared to companies similar to you. We’re also looking to open up our AI platform by giving you the capability to bring in ML applications and models, and leverage all of that on your Anaplan platform. Now, based on some – this morning I was having a conversation with one of our customers, and it appeared to me that sometimes there’s a perception that PlanIQ, Optimizer, Predictive Insights are all separate products. But basically what it is, it’s all on your single Anaplan platform, where you’re able to bring in internal, leverage your internal data, but also leverage external data, combine that with your statistical algorithms that’s currently available on the platform, but also leverage tools like Optimizer, Predictive Insights and PlanIQ, where you could leverage even more advanced ML models. You do not have to be a data scientist to figure out if you should be using Prophet or any other one of our ML algorithms. PlanIQ could do all of that for you, where it selects the best algorithm. You have the capability if you want to go and select a specific algorithm, but not many of us are data scientists. Let the machine do that work for you.

Lesley Vincent 0:15:16.8:
Future engineering, you don’t need to hire a data science team. If you have a data science team, that’s okay, you could leverage them, but PlanIQ could automatically do your future engineering for you. I came from a world, before Anaplan, where you will basically stitch different AI components together, and we call it an AI solution. What that means, you need to pay a partner in perpetuity to manage and scale that for you. Our strategy at Anaplan is one single platform, capabilities that are powerful enough for your data science team to use, but simple enough for someone from the business side to implement. If you look on our website, you will see a testimonial from a woman, [?Rene Taylor] from Coca-Cola, talking about how easy it was for her to implement PlanIQ and scale that. She’s not a data scientist, she’s a business user on the finance side, she was able to do it quite easily. Then on the other end of that, the business outcomes, we see an improvement in your forecasting accuracy, but we don’t stop there. Deeper insights, these tools give you deeper insights into your business and it also improves the efficiency of your processes.

Lesley Vincent 0:16:44.6:
What I have seen, what we have seen, our customers coming to us and saying we want to improve our accuracy, but we also want to make our processes more efficient. Now, before implementing PlanIQ, or any of your AI/ML solutions in your organizations, there are six critical factors that you should consider. Obviously, very important is number one. Your executive sponsor should be actively engaged in the implementation of the project. I don’t mean hands on keyboard, obviously, but if you look at number one and number six, these are connected, so with your executive sponsor actively engaged, that executive sponsor is now able to help drive change management, which is absolutely critical in any AI/ML implementation. Without those, it will be very difficult for you, even if you have the most advanced AI/ML solution that you’re trying to implement. Then number two, define your goals and success criteria. Obviously, absolutely very important. If you have an accuracy goal that you’re trying to reach, that should be well defined and tracked. They’d have access to the right data, having the right amount of data, two years, three years, the right quality of data, very critical for your project.

Lesley Vincent 0:18:18.2:
From a Center of Excellence perspective, the methodology that we use to implement our projects is called The Anaplan Way, so ensuring your team. There’s a couple of hours training on our website on The Anaplan Way where your team could go in there and just get updated on that iterative methodology to implement any AI/ML solutions within your business. Also very important, ensure that a subject matter expert from your organization is embedded into this project, someone that understands your business, how it works, understand how to measure and track the success criteria, able to talk to the business, etc. These are the critical success factors that you should consider before you even lay hands on keyboard. Then we continue to work with our team, we continue to learn from the implementations that we have done, we continue to work with our partners to think of a data strategy when you’re implementing AI/ML. The three main categories of data that we see are industry related data, so thinking about customer insights, vendor data, competitor data from sources such as Nielsen, IRI, social data, community data, so your weather data, property data, from Twitter or Facebook, and then your financial and regulatory data from Bloomberg, LexisNexis, etc., and then stepping back and contextualizing that as it relates to your use case.

Lesley Vincent 0:20:00.3:
So if you’d be implementing demand planning, revenue planning, OpEx planning, you have to think about one, which of these data sets that you’re already subscribed to, and then also think about which of these data sets will have the biggest impact on day one of your project implementation. Instead of trying to think of boiling the entire ocean, looking at the data sets that would have the impact today, and then work with your business partner and your implementation partner and your internal team to think about a road map. How are you going to implement, incorporate these additional data sets and test them and ensure that they’re actually impacting your focus or whatever AI/ML solution, Predictive Insights that you have implemented. From my implementation strategy, what we’ve seen basically are just different configurations that you could stitch together to implement your AI/ML solution. You could go with Anaplan professional services. Basically, what that is, is a team of some of our best solution architects and master Anaplanners that could be embedded in your organization and work with you to implement any AI/ML solution that you’re looking to implement.

Lesley Vincent 0:21:21.2:
Partners, we have a well-developed and mature partner, implementation partner ecosystem. That’s another option. You could bring in a partner to drive that implementation for you. Some of our customers have very well-developed Center of Excellence, your internal solution architects and model builders, PlanIQ, Optimizer, these tools are incredibly easy to implement. Two years ago, it took me just a couple of hours to implement PlanIQ, and the reason why it took a couple of hours, because I need to actually run the focus, have the algorithms do their thing. It’s just three screens, click button, it’s very easy to implement. We have training that you could go through. We have trainings for all of these products that you could go through and implement it. You could also work with a partner or professional services team, so you could have a configuration of that, your internal model builders plus professional services. Now, during implementation when Adam introduced me – I’m a Customer Success Business Partner. I’m also the Industry Captain, but I would also rely on your business partner to participate in your project meetings, facilitate success reviews with your executive sponsor, keep them involved, keep them updated, and then enable your internal team with training.

Lesley Vincent 0:22:53.4:
We have partners who are capable of training you on these tools, but we also have an online training, what we call Anaplan Academy, where you could go and do these trainings and quickly be skilled enough on these tools. Then obviously your business partner will liaise with your implementation partner and the internal team, so usually I’m talking with customers from our technology manufacturer and different industries, so business partners typically bring me in. If it’s another product, they will bring someone like a solution consulting consultant in to work with you, to talk to your executive sponsor. If you’re working on a deal or something, that business partner could liaise and bring in other folks from Anaplan. Now, from an implementation perspective, you just heard me say that these tools are incredibly easy. I’m talking about plain vanilla, PlanIQ, Optimizer and Predictive Insights. It will take you less than a day to implement these things, but as we know, business, real world business is obviously more complex, so what we’ve seen from an Optimizer perspective at Autodesk and one of our automobile fuel retailer, three months of implementation.

Lesley Vincent 0:24:14.9:
From a Predictive Insights perspective, one month or five months at CDW, and at Coca-Cola, that was a one-month implementation. At one of the largest retail banks in the world was a five-month implementation. The variation of the time is how complex your requirements is, what problems you’re trying to solve, etc. I wanted to share this feedback with you from Autodesk. Basically, Autodesk had a Python solution that they were using to figure out their go-to-market strategy, so if you think of it on an annual basis, your go-to-market strategy will change and any change that was a manual process, where they had to configure that Python application, that was incredibly difficult. No one wants to run the business that way, so by leveraging Optimizer, that process was automated, much more seamless, and they were able to basically analyze and optimize their marketing, the go-to-market strategy for millions of their accounts using Optimizer. After Go-Live, your business partner should not, would not, will not go away, so on a quarterly basis, your business partner will have, what we have quarterly success reviews.

Lesley Vincent 0:25:38.6:
Please try your best to ensure your executive sponsor remain engaged because they’re paying for it, they have to sit with your C-level or your board and justify their investment in this capability, so your business partner should be the one facilitating this quarterly success review meetings. Bring your executive sponsor and have that conversation. Recently we launched what we call a Verified Outcomes program. What that means is the success criteria that you define, your targeted business outcomes that you define at the beginning of the project. We will, your business partner will have a one-pager where you’re tracking that and presenting that to your executive sponsor, so remember that verified outcomes or an executive summary of your business outcomes that you’re targeting for your AI/ML. That way your business sponsor is engaged, you will have your eyes on the value and the business outcomes that you’re trying to obtain with your AI/ML. Your Connected Planning journey, basically what that is, your business partner, implementation partner, your internal team will continue to work with you to find additional or other areas within your business where you could get more value from our AI/ML offering. You could start today with Optimizer with your linear programming use case. Oh, I’m trying to increase revenue in X, Y, Z, then you move on to Predictive Insights, PlanIQ, and then as Adam mentioned, you implement an app.

Lesley Vincent 0:27:14.5:
You could leverage all Natural Language Querying, so it’s by working with your business partner you could actually figure out this additional areas you could bring more value. Then here in connection to that, your business partner obviously will be closer to some of the enhancements that we continue to make in the product. Let’s take PlanIQ, for example. Most AI/ML solutions, they will have the Prophet algorithm, but if you look at PlanIQ, you will see two versions of Prophet. The first version of Prophet is the plain vanilla Prophet that was created by Facebook many years ago, that’s used for time series, forecastings, use cases such as demand planning, revenue planning, etc., but we have Prophet 2.0. What that means is that algorithm, plain vanilla algorithm now has explainability, so by working with your CSBP, or your Customer Success Business Partner, that individual will let you know these enhancements and how you could leverage them within your business. Then continue to support your team, but also very importantly, just expect your business partner to leverage people like me and other folks who knows the AI/ML really well, to just help you think about, hey, additional data sources that you could bring in to make you, to increase the accuracy of your specific AI/ML use case. With that, I would like to open up Viji and I for any questions that you may have.

Viji Doraiswamy 0:28:54.4:
[?Katie] has a mic in the back, so it’d be good to…

Unknown speaker 0:28:56.9:
[Unclear words 0:28:56.9].

Viji Doraiswamy 0:28:57.8:
Yes.

Unknown speaker 0:28:58.3:
Oh my gosh, great.

Audience 0:29:01.3:
Hi, thank you very much. Really kind of curious on the external data part on – I saw you have competitor data, so where is that pulled from? Is that just going to be kind of like public information-type things, or is it going to be based on trends that you have from within the app already? If you could just expand on that a little bit?

Lesley Vincent 0:29:19.6:
Great question. We are obviously looking to become more mature on that, and Adam talked about data management system this morning. Currently, you have to bring in the data, bring it into one of your modules. You could drop it on one of your cloud folders etc., and we would use a tool called [?CloudWorks 0:29:39.4] to bring that data into PlanIQ, for example. It’s all – some of our customers already have subscriptions to IRI or whatever, so we – that’s the obvious place to start, and we will work with you, we’ll work with your implementation partner. As a business partner also, we have visibility as to what other customers are doing and where the industry is going, so we’re not going to share use case information with you because that’s people’s, customers’ private information, but just the trends. We will say, oh, for demand planning, we think weather data makes sense, etc., holidays. [Unclear words 0:30:23.1], calendar, data holidays are already built into PlanIQ, so the click of a button you could leverage that.

Audience 0:30:40.0:
I want to know if there are any plans to connect the [?ALM 0:30:43.7] concept to management reporting to dynamically build a deck based off of a query?

Viji Doraiswamy 0:30:52.6:
That’s a use case. That’s new to us, so we haven’t – I don’t think we’ve considered that use case, but happy to put that out on our list, as we are still in the pilot phases for some of these.

Lesley Vincent 0:31:05.9:
It’s a great idea.

Audience 0:31:14.1:
Hi, could you guys also talk a little bit about the technology that you guys are using for the [?Gen AIPs 0:31:18.1], like, what are the models underlying, and did you guys do any fine-tuning on top of that, or how does that work in the context of the enterprise?

Lesley Vincent 0:31:27.7:
Yes, as it relates to whether we use what foundational models and how we’re fine-tuning, and how we’re building that information, unfortunately I wouldn’t have. That would come from our product team because they – currently the development and development phase of that, yes.

Viji Doraiswamy 0:31:46.0:
Adam talked about our using OpenAI for sure, so we are looking at that partnership with OpenAI, but yes, the models etc., are still under, being revised and developed.

Audience 0:32:00.0:
Hi. You mentioned briefly about third party integration of machine learning models. Could you share a bit more on that, especially for customers who do have a mature data science team in-house?

Lesley Vincent 0:32:11.3:
Yes, well, I could – so you’re currently… PlanIQ, for example, you could leverage, if you were using a third-party tool, you could leverage that, but what we’re looking to do in the near future is to make that more seamless, simplify that, yes. You have the capability to do that today, but we’re on our road map, we’re looking to simplify that process, yes.

Audience 0:32:43.7:
When you deal with your customers, do you have a preference if they develop their own models training on open source, or just use the model infrastructure of private proprietary infrastructure like OpenAI? Do you have a preference?

Viji Doraiswamy 0:33:04.0:
The road map, as part of the road map, eventually, based on what he said and what you’re saying, we do want to make sure that customers are able to bring their own data models, their own machine learning algorithms, and make our platform more open. I think at this point, as we are developing the journey, we are working with what is – we expect you to work with what we have, but eventually down the road, we do want to make it more open for you to be able to bring your own models and algorithms. Does this excite you all to start going and building a business case with your leaders, to start at least developing some of these AI pilots and use cases?

Lesley Vincent 0:33:52.3:
I’m always available, by the way. Talk to your business partner, happy to get on a call with you to talk more about the use cases that you’re implementing and how we could bring it on Anaplan. [Laughs]

Unknown speaker 0:34:08.0:
If there’s no further questions, thank you both to Viji and Lesley.