SuperAnalyst: Move from Intelligence to Execution, Around the Clock | AlphaSense

0:04 Welcome everyone. on behalf of AlphaSense, I want to thank you all for joining us today for our webinar super analysts move from intelligence to execution around the clock. I'm Sammy Esman, a product marketing manager here at Alphasense and I am really excited that so many of our customers are joining us today and want to extend a warm welcome to those of you who are new to AlphaSense. Before we begin, I'd like to share a couple housekeeping items. On the bottom of your Zoom screen, you'll see a tab for some resources related to Super Analyst. We also included a link to our 2026 Alpha Summit, which is happening this October and where Shaquille O'Neal will be on the main stage. We're offering webinar attendees $100 off the registration prize. So, be sure to check that out. Also, on the bottom of your screen, you'll see a Q&A panel. You can send in your questions at any time during the presentation and demo and we will get to the questions afterwards.

1:01 we will try to answer as many questions as we can during that time. And with that, I will turn it over to Chris. >> Thanks so much, Sammy. and thank you so much to everyone for joining. We are so happy to have you all here and we are incredibly excited to tell you more about Super Analyst. We've got a packed session today, so we'll dive in. my name is Chris Aerson. I'm senior vice president of product at AlphaSense and I'm joined by Louis Boa, our director of product for generative AI. I'll start by setting the stage with a bit of context before I pass it over to Louie who will show you what superanalyst actually looks like in practice, the fun part, and we'll save some time at the end for Q&A. So, please submit your questions as Sammy said. Let's get into it.

1:48 So when we started at AlphaSence, we set out to solve the problems we were seeing with research. Financial and strategic professionals were drowning in documents, broker research, earnings transcripts, filings, expert calls, and finding the signal that mattered was a full-time job on its own. Professionals were spending time finding the right documents, more time finding than they were making the decisions that really mattered. Smart synonyms was our answer. It was the foundational semantic search technology that cemented AlphaSense as purpose-built for this work. A search engine that really understood language the way a practitioner does. That when management talks about layoffs, they may say workforce reduction, job cuts, or use even more coded language like reorganization or flattening.

2:39 For the first time, you could eliminate blind spots and make confident, highstakes decisions without spending hours trying to find the right document. Then technology and the problem evolved. Finding information wasn't enough. You needed to make sense of it fast. So, we built generative search. Generative search went beyond surfacing a list of documents and started to actually synthesize them at scale. We compressed hours of reading into minutes of insight, always cited back to the original source. That was the moment of stopped being just a research tool and became an intelligence layer that took complex research and made it simple.

3:23 With deep research mode, that time savings became days or even weeks as we introduced a gentic AI that could reason over larger sets of documents, hundreds of documents, run searches on its own, and make decisions about what to do next. And we didn't stop there. Last year, we launched workflow agents, the connective tissue between intelligence and action. We gave you the ability to take those research capabilities and automate them, monitor a portfolio, track a competitor, run a recurring workflow, all with a single click. In that move moment, we moved from answering questions to executing tasks and replaced the manual repetitive work that was getting in the way of high value strategy.

4:14 But here's what we kept hearing from our customers and what really keeps us motivated to to continue to innovate. The workflows still stopped when you stopped. The intelligence was only as current as the last scheduled run of a workflow agent or the last time someone was motivated to run a search. The gap between insight and finished deliverable was still being bridged manually by your team. And that is the gap we're closing today with Super Analyst.

4:45 Autonomous execution that scales every team member's ability to monitor, analyze, and deliver decision ready work faster. A teammate, not just a tool. Best-in-class content, generative AI, automations, persistent memory, and deliverable creation 24 hours a day, 7 days a week. An agent that works while you sleep.

5:15 So now you have the context on our journey from search to synthesis to execution. Each step built on the last, each one made possible because of the trusted foundation underneath it. But here's something worth pausing on before we get into super analyst. The market right now is moving really fast. There are a lot of agentic AI tools claiming they can do what you just saw in that timeline. The assumption embedded in most of them is that connecting a model to more sources is the same as getting better answers. The truth is that it isn't. Access alone is not intelligence. And for teams making investment decisions, running competitive diligence or advising boards, that gap has real consequences.

6:03 So what makes AlphaSense different? Generalist AI tools, they they generalist AI gives you the tools but expects you to separately license the sources you need. AlphaSense's foundation is best-in-class content. We're talking about 500 million plus premium business financial business and financial documents across broker research, expert transcripts, filings, earnings calls, and your proprietary internal content. Much of this content you can't find anywhere else. But the question isn't just whether your AI can connect to a source, it's whether it will actually find the right evidence and know which evidence matters most.

6:46 AlphaSense knows which sources matter before the model ever sees them. That's because of the knowledge graph we've been building for the past decade and a half. Every day, hundreds of thousands of documents are ingested, enriched, tagged, connected, and truly understood by our models. Now, impressively formatted outputs can mask weak context and missed evidence. AlphaSense's decision-grade AI doesn't just look right, it can prove it. Every answer is traceable. Not just a citation at the bottom of the page. Every output is cited and fully auditable in line in seconds. In the work most of you are doing, the difference between an output you can defend and one you can't is often the difference between adopting it in real work or not.

7:39 And ultimately, none of that is helpful without considering efficiency. Most MCP connections hand all of the heavy lifting to the LLM and hope it has the skills to find the signal and brute force its way through a fire hose of content. Alvisense intelligently narrows the context before it ever reaches the model, which means higher quality outputs and none of the overhead that comes from handcuffing a Frontier LLM with second rate search and retrieval.

8:13 Every product we build, including Super Analyst, inherits this same intelligence layer. Our AI tools are good because of what's underneath them. Super Analyst is a power user of AlphaSense just like many of you and that's what makes everything you're about to see possible. Now the market has been telling us something over the last couple of years. According to McKenzie's 2025 state of AI survey, 62% of organizations were at least experimenting with agents and 23% were actively scaling. While Gartner projected that 40% of enterprise apps will include integrated task specific agents by the end of 2026, which is up from just 5% in 2025.

9:00 And what we've seen in practice is the work itself is changing. The expectation, your expectation has shifted from AI that assists to AI that executes. from tools that are fast to automatic, from systems that require prompting to systems that are autonomous. And the teams winning right now are the ones who found a way to close the gap between intelligence and execution. That's exactly what Jack, our CEO, described in the superanalyst announcement.

9:32 Today's decision makers are overwhelmed not just by information itself but by the sheer volume of manual work required to turn information into decisions. Super Analyst becomes an extension of their teams by continuously monitoring, analyzing and completing workflows in the background. This isn't about a single feature announcement. It's the difference between a tool you operate and a tool that operates on your behalf. And it's a different way of thinking about what a platform like AlphaSense can be.

10:06 I want to spend just a few minutes on the capabilities, not to walk through a feature list, but to give you the right frame for what you're about to see. The way I think about Super Analyst is this, not as a tool, but as a power user of the platform. Super Analyst is different because it has access to all of our inst in institutional grade tools and data and because it's proactive. It's always running, always monitoring, always advancing the work at your direction whether you're in the platform or not.

10:41 There are really four things that make that possible. So the first is always on execution. So it never misses a signal. Superanalyst doesn't wait to be asked because all of this content and data flows through AlphaSense every day. Superanalyst monitors filings, earnings, transcripts, and market developments continuously. The moment something relevant happens, it moves. This means your workflows aren't dependent on someone remembering to run them or even you having to remember to set the alerts and workflow agents to run at the right times. They just run.

11:17 Second is persistent memory which means super analyst gets smarter every session just like an analyst you are working with on your team. Every project you work on superanalyst builds on itself over time. The agent knows your mandate, your methodology, your preferred outputs and the work that came before. You're never starting from scratch. You never have to reexplain the context. You just manage and direct. every session picks up exactly where the last one ended and over time that really really compounds.

11:54 Third is decision ready outputs. This is the one I really want you to sit with. Super analyst doesn't produce just chat just answers in a chat window. It produces finished work products, briefs, earnings previews, financial models, slide decks, memos, cited, structured, and ready to act on. The goal isn't to give you more to read. It's to give you less to do. Super Analyst is purpose-built for your workflows.

12:24 it's preloaded with out-of-the-box skills, which means it doesn't require you to build anything from scratch to get started, although they're really easy to customize. These skills are reusable prepackaged workflows that the agent already knows how to run. Things like company deep dives, key debate analyses, pitch decks, and Excel workbooks. So when you open Super Analyst for the first time, you're not starting with a blank environment. You're already starting from someone that works understands your industry.

12:56 Our vision is a world where every professional on your team has access to something that works like the most cap capable analyst you've ever hired. One that never sleeps, never loses context, and gets better the more you work with it. That's what we're building towards and what you'll see the early innings of today. Now, before Louie gets into the demo, one thing I want to address headon is security because we know it's the first question that a lot of teams have when they hear the agent is always running in the background or it has persistent memory. So, let's get the most important thing out of the way. Your data is yours. Super Analyst runs in a private per user environment. Nothing you upload, generate, or store is accessible to anyone else. Not other members of your team unless you choose to share it, and definitely not other clients. Your workspace is isolated from the moment you create it.

13:59 The persistent memory piece is something I want to be especially clear on because we've heard the question, if the agent remembers my preferences, does that mean my data is training the model? And the answer is unequivocally no. Super Analyst writes your preferences and workflows into a private file system that only your instance can ever access. That data never feeds the underlying model and it never benefits another user. And for those of you thinking about security posture more broadly, Super Analyst runs on the same security infrastructure that governs everything on the on the Offsence platform today.

14:36 It only accesses what you explicitly give it access to. Everything stays within the platform. We'll have more detail on the architecture available for your IT and security teams and follow-up webinars focused exclusively on security. The short version for now is this was built to meet the standards of the most security conscious organizations in the world. Now with that, let's get to the good stuff. let's show you what Super Analyst actually does. Over to you, Louie.

15:08 >> Awesome. Thanks, Chris. hey everyone. I'm very excited to demo a super analyst to you all today. as a quick introduction, I'm Louie, director of product at Alpha Sense. I've worked on our flagship generative search and deep research products and I'm privileged to be building, you know, super analyst with a world-class team here at AlphaSense. I'm going to start by setting the stage and sharing my my mental model before before sharing my screen. so just to kind of reiterate what Chris said, so the way that we think about Super Analyst and I'd encourage you all to carry this mental model through the whole demo is that, you know, Super Analyst is your personal sense power user. it knows the platform inside and out, including the content, the tools, the metadata. It has skills that match your most common workflows. it can code to put together long complex workflows or to create beautiful artifacts. and it has persistent memory which means that over time it starts to know the way that you work and what your preferences are. I'm going to go ahead and share my screen now.

16:21 there we go. so just a quick note that you know as we are still in development some of the experience may still change and I'm going to touch on some of the upcoming road map items at the end of the demo. So here we are on the superanalyst homepage. so super analyst is a technology layer that we've built into alpha sense. So the entire experience should feel very familiar to those of you who have used generative search. you are going to work and talk with the super analyst in natural language very much the same way that you are talking to a member of your team. okay. So, let me start by setting the stage for our demo today.

17:02 so company ramping is something that we all do, whether it is to get up to speed on a new name, update a thesis, do competitor research, or to benchmark against. and for this demo, we're going to pretend that we need to quickly ramp up on on Uber. So, the next print for Uber is roughly 6 weeks out, and Super Analyst is going to help us get from zero to fully briefed. So the scenario here is maybe you're a new analyst picking it up or a corporate strategist benchmarking your own business against it or an IR team at a competitor trying to understand how the street is thinking. So the following three questions are going to follow you for the next 6 weeks. So firstly what are my expectations for Uber for the next quarter based on what is already in my model and then what is the signal from experts that I should overlay on top of this. The second one is what are investors genuinely split on right now.

17:58 So where's the real key debate among investors? And then third is how do my independent estimates stack up against the streets estimates and guidance. Okay, so let's begin. I'm going to paste in my first prompt here. So, the first prompt we're going to start off with is forecast the next quarter's results for Uber based on historical trends and expert signals. So, I'm going to go ahead and hit run.

18:29 So, the agent is starting to work. So, this will run for a couple minutes. So, in the interest of time, I'm clicking over to show you a thread that I already ran in preparation for this for this demo. so so let's take a look at what actually happened here. So here you can see the tool calls and reasoning stream from the super analyst for the specific query. So if this were to run live, these will kind of follow live on on on your screen and you will be kept up to date by the super analyst on the tools it's using and the reasoning and the decision it's making as it's working to give you the final answer. So firstly we can see here that super analyst told me that it has a very current V4 forecast from June 22nd.

19:12 you know that is demonstrating some of that persistent memory. I'm going to touch on that a whole lot more in a bit, but that's not surprising because I've been preparing for this demo. and I ran some of these queries yesterday as well. So then here, it starts pulling in some skills as well. So I can see here it says good all skills loaded. I can expand this to see some of the skills that the superanist query the roll forward skill, signal overlay skill, build artifact, and driver tree skill. I'll touch more on skills in a minute. And then it starts getting to work. So it starts pulling the candalyst model. It simultaneously pulls fresh expert signals. it extracts data from our canalist financial model for Uber.

19:52 It pulls some of the mobility and delivery segment data. you know it does a fresh broker signal sweep and then it starts building the artifact. So I'm being kept in the loop by the super analyst while it's busy doing its work. So at a high level just to kind of set the stage for what is happening in the specific query. There's kind of two parts to it. The first part is the super analyst takes an Excel model. In this case, I selected to use our catalyst Uber financial model. You will be able to provide your own Excel model and then it takes the historical data inside of that model for the key KPIs and drivers and it rolls that forward. So that's pure mechanics. It's based on the trends and the growth and the seasonality inside of that Excel model. Then the second thing that it does is it then looks at what experts are saying in our corpus of data inside of Alpha Sense and it creates what we're calling an expert nudge. It essentially looks at the different signals from experts and brokers and it creates an expert nudge on top of that roll forward number for the next quarter. So I'm going to scroll down here. Here it starts to kind of stream a summary for me in the in chat.

21:00 And here of course everything is is is cited to the source. So I can click on this expert call you know it opens up in chat where I can see that this specific expert is saying you know the company faces significant risk from a technology providers like Whimo which sources to the document inside of AlphaSense. but I'm going to go ahead and show you the artifact. So it gives me two outputs. One being this dashboard file and then the other one being an Excel model. We're going to start by opening up the dashboard.

21:36 So for this specific example, Super Analyst built a fully interactive dashboard in just a couple minutes and it includes five tabs. This is fully interactive. I'll click through the tabs really quickly and then we can kind of talk about what is in each one. So the overview tab here is my headline summary. It includes keymetrics table with actual history on the left, roll forward in the middle, and the super analyst estimate on the right. It also shows a two-year stack sanity check and key expert factors that drove each estimate. The roll forward is that purely mechanical roll forward number based on the historical trends and seasonality inside of the Excel model.

22:16 It goes into detail on how it actually does the calculations and the methodology that uses so that you can see for every single KPI the calculations that were done in order to roll that forward into next quarter forecasting. This is where you can see where the specific signals are coming from. So where is the super analyst finding signals in our corpus of data that it applies as an overlay on top of the specific role forward numbers. So in this case you can see the factor summary KBI tables. You can see where it found evidence, the number of sources it found, the net expert nudge that it does and then essentially the adjusted driver. Again, all of this is sourced to the underlying documents inside of AlphaSense.

22:59 The model tab is just an historical look back at the model that you provided and then sources. This is the data provenence list that pulls together all of the sources that form part of this super analyst forecast skill. So I'm going to close the file here and just show you all of this is also available in Excel. so I can view the Excel inside of the viewer inside of the super analyst. I can also download the Excel file to my machine and then open it up. so here's the overview tab, the model, and then signals and evidence. And then here again, here's the kind of verbatim quotes that speak to where the super analyst found specific signals that is that it is using for that overlay.

23:55 Okay. So, let's now take a deeper look at some of the mechanics here. So, I'm going to go and scroll back up and actually open up this first item here. So, this is where you can see under these work, this is where you can see where the super analyst is doing work and what tool calls it's making. So, here I can see that the super analyst has read this super analyst forecast skill. And I want to spend some time on skills because it's such an important part of essentially the power of the super analyst. So, I'm going to go over to the skills tab.

24:26 where I can show you some more of the skills that we have. So firstly, for those of you who might not be familiar with the term, a skill is essentially a package of instructions that tells the agent what to do. So it can include a prompt, it can include templates, reference files, and even scripts. it is phenomen phenomenally powerful, and it is also very simple to use. So when you start out with your super analyst, it's going to come preloaded with AlphaSense workflow skills. So you can see some of the alpha science workflow skills in the section that I'm scrolling through already. So I've set up my super analyst to be a hedge fun analyst. So the skills that you're seeing on my screen are heavily focused on financial services workflows. but we are building out a suite of ski skills for every role across financial services, corporate and consulting jobs to make it very easy for you to get started with your most common workflows.

25:23 and then the real value in my opinion is just how easy it is to build your own custom skills. So here you can see some examples of custom skills that I've created. All you have to do is ask in natural language. and our skilled builder, which knows the ins and the outs of the AlphaSense platform, will take you through exactly how to get the most out of Alpha Sense to build your specific use case. And then once you've set up a custom skill, it is yours and only yours and you can run it whenever you want to.

25:54 Okay, so let's let's pause here. so now we have an initial number from the Superranos Force gas skill, but we all know that a number alone isn't conviction. So I'm going to click over to our second example for today. Before you put capital behind a view or before you walk into a board meeting with a competitive benchmark, you need to know what the market is arguing about. So what are investors most divided on for this company right now? So here we can see the output of another super analyst skill that I'm personally very excited about. This is called the key debate skill. Now this is doing something very very different. It's not asking what everyone thinks. It is looking for real divergence amongst brokers, experts and management. So, it reads through broker research notes, expert call transcripts, management commentary, and finds the questions where the bulls and the bears are most split because that's where the alpha is. And for a corporate strategist, that's where the market narrative risk lives. Here you can see the current four top key debates. I'm going to go ahead and open this into widescreen.

27:08 So, not only does the superanalyst surface the four questions that matter most, but for each of these key debates, super analyst shows you where every broker, every expert, and management itself stands on each one. I'm going to go ahead now and click through this dashboard to give you an overview. So, here on the overview page, we can see a summarized table of the key debates for Uber as well as some of the risks to watch. Again, everything is cited back to the source. Each one of these key debates has an individual page that tells you what is the debate, why does it matter, what decides it, and then goes into a bull versus bare view across experts, brokers, and management. So, I can in this one dashboard look at the top four questions currently where investors are mostly arguing about what is important and what are key drivers for Uber. And I can see this individual bullbear debate for for Uber. So I'm going to click here delivery durability versus door dash a disruption. And then here's all the sources. So here's the provenence log for the key debates dashboard where I can see all of the sources used inside of this keybait skill. So this is where the real edge is. It's seeing where the three sources diverge because when an expert who left the company 6 months ago contradicts what management said on the earnings call, that's alpha.

28:35 so I'm going to go ahead and show you something else that I think is really powerful about the super analyst. So again here by going into seeing what is happening when I ask these queries I can see that here the super analyst actually search searched for memory. So this is what Chris touched on. so super analyst has persistent memory and what this means and I'll open the memory file here for you.

29:06 I can just kind of shrink this a little bit. This means that as you work with it, it'll start remembering things about you to make it more intelligent at helping you. So, it remembers things like your preferences, the industry you cover, and companies you researched. What I'm showing you on my screen right now is my memory file. So, let's pause there for a moment because I think that's important. This file that you're seeing in my superanist environment and it's only accessible to my superanist. It is not training the agent. it is just there to search and update. The way that I think about it, I'd encourage you to do the same, is it's exactly the same as taking notes and then referring back to those notes notes later to remember what was said in a specific meeting or conversation.

29:54 But what this means is that as you work more with your super analyst, it'll remember what you like, don't like, it'll know how you want to see your answers displayed, documents formatted, the last time you asked about a specific company, and where to pick research back up. It'll become familiar with the way you work and what is important to you. The exact same way a new starter on your team would learn the ropes as they onboard. So I can scroll through my memory file here. So here's some of the research notes that I can see.

30:30 You can obviously see there's a lot on Uber from the last two days. And then couple other companies that I've done. And then here's some of the user preferences for me where you can see some of my preference around communication style, formatting, my role. okay. There was a question in the chat, maybe we can just answer it here, which was, can you edit that file? And yes, absolutely. That that's your your files that you have control of over. So, so you can go in and edit them directly.

31:04 so that was a great question. >> Awesome. Thanks, Chris. Okay, so let's take stock. So, we already have a model with a prediction for the next quarter, and we now have the key debates mapped. So, there's one more thing that we need, and this is a complete picture of what the street is expecting sector by sector, line by line. So, I'm going to go ahead and click over to our third example here, which is the earnings preview skill. So, what you're looking at on my screen right now, is the output of the earnings preview skill. And this is your entire preprint package. it synthesizes everything. Estimates, broker positioning, the key swing metrics, the exact key debate questions that will drive the stock on earnings day. you should think about this as your earnings day cheat sheet.

31:58 What used to take you a couple days to produce, super analyst creates this first draft in a couple minutes. you can simply continue working in Excel or ask the superanis to change anything in the output. you're still going to want to review the source materials, read the broker notes from your favorite analysts, and Super Analyst helps you organize them for easy consumption. This is how easy it is to go from research to first draft deliverable with Super Analyst. Let's click through the dashboard and Excel together. So, just a quick kind of click through here. So, here's my earnings preview skill for Uber. I can see the key drivers with the sources linked to where those drivers are. The expectations table positioning key debates this quarter bullbear base case and then the references used.

32:53 Again, this is also created in in this case, it creates another artifact which is a more report style view that you can see in the doc view that you can download as a PDF and it also creates a Excel file for me. I can scroll and look at this in the view. I can also download this to my local machine as we saw. Okay. And lastly, I want to end with a bit of a tangential example to to demonstrate how super analyst acts like a power user and has and has AlphaSense command. So I'm going to go ahead to this example.

33:31 So for those of you who are not yet familiar with the AlphaSense platform, this will also give you a glimpse of the existing platform. Because Super Analyst is built into AlphaSense, it has access to our vertically integrated platform. This means it has access to our content, metadata, tools, and workflows to work the same way that you would work when accessing the platform. So, here's an example where I ask the superanas a couple questions and I'm going to click through the answer as well as where the information is on AlphaSense to show you how the superanis works and navigates across Alpha Sense. So here in this case I asked give me a list of ride sharing comps for Uber. So here we can see that it read this comp's identification skill. It started looking for data.

34:21 It looked for data in some of the filings as well as in the candalyst industry comps. It then starts streaming the answer in chat. It splits between best comps, public comps, adjacent multimodal, crossindustry, private comps. So I followed up by saying create a watch list for me with the ride sharing companies that are on the industry coms. So I'm going to go ahead and kind of demonstrate what the super run has done here. So this is industry comps. if I go ahead in here and I filter for Uber, I will see the industry comps that Uber exists on. So in this case, what we asked for is we asked for the ride sharing industry comp. So it essentially pulled this list of companies for for from this industry comps list. So this is a full comps table with industry KPIs and KPIs for these companies that the super analyst has access to.

35:14 I then followed up by asking let me go into this one. I then followed up by asking to create a a watch list for me. let me just scroll a little bit up to find my spot. So create a watch list for me with the right sharing companies that are on the industry com. So here we can see that it worked. It read this watch list skill because the super analyst is integrated into AlphaSense. It can create watch lists that I can then use in other places on the platform as well. So to demonstrate that I can head over here to the dashboards tab, go to my watch list.

35:53 Here's my ride sharing comps watch list that was created today. Now I can go and manage that from here. And here's the watch list that the super analyst created for me in that specific request. So this is demonstrating that that full offense command from from the super analyst and then following up from the watch this for me. I asked kind of two interesting and exciting questions. I asked the first one which was now go and download the candalyst financial model for each of these companies and then create an interactive time series chart for me to see mobility gross bookings over time for the past eight quarters across all these companies. So here you can see it started working. It downloaded all of the financial models for these different companies. It started parsing the data. It found the correct metrics from each of these companies. It found the eight quarters of data and it starts creating my chart for me. So here I can see my time series chart. I can expand this a little bit to make it more clear. So this is again configurable between line and bar. It has the data inside. It has the specific metrics here. And this is all now pulling from from the underlying canvas models that the superanis created. And then I followed up and this is the last example here. where I now said okay now let's look at some of the qualitative data. So I said looks great.

37:14 now please create another tab in this dashboard where you pull expert and broker quotes with specific comments on ability growth bookings estimates and growth outlook. So now I'm trying to combine the quant and the quell with the super analyst. So it starts firing these searches in parallel. It eventually finds some rich data here, excellent data. I have everything I need and then it starts creating this tab for me. So now I have this view where it has the time series chart and it has the broker and expert intelligence in here.

37:42 So now I have these different quotes from brokers and experts on the specific metric for all of these companies in my watch list. So now I can scroll through here where I can essentially see all of the different quotes that it was found in this information. Okay. yeah, so I think this is a great example of how you would work with the super analyst to explore the AlphaSense platform. you know, once you end up with something like this that you like, you can also just ask the super analyst to convert it into a skill and you can reuse it again later. That's a workflow that I've used a lot. I work with the super analyst. I find out that that's something that I like and then I ask the super analyst just please convert this into a skill for me and it'll look back at all of the steps you took to create that skill for you. Okay. Okay. So, I'm going to end my screen share here and just do a quick recap before we get over to to Q&A. So, we just saw Super Analyst answer three important questions to help us understand Uber as a business and prepare us for the upcoming print. So, in order to do this, to recap, Super Analyst used the AlphaSense tools just like a power user would. It downloaded the canalist models. It searched across broker documents, expert transcripts, company documents. It used skills to run recurring workflow stepby-step instructions and execution. It used its persistent memory to look for previous research on Uber. And it made sure that it picks up where it left off previously. And it built these interactive dashboards and Excel models for me to use asis or to further customize with the super analyst. we also saw how the super analyst can use the AlphaSense platform just like a power user would by creating a watch list, finding industry comps, downloading models. So the things that are in the demo today, persistent memory skills, the ability to create your own skills, the interactive dashboards, the ability to build Excel models, those are not going anywhere. And then looking at the road map, we're still adding more capabilities. Things like always on automation, document upload capabilities, event triggers, live updating dashboards, additional access to tools on the AlphaSense platform, connector support, and deeper integrations with AlphaSense for PowerPoint and Excel. Our goal with automation is to be able to take everything I just walked through and then to ask superanalyst, run it all automatically before the next and every future earnings. You only set it up once, but your super analyst will continue working based on the triggers that you set. As I said at the beginning of the call, we're still in development, but the first big milestone for us is getting clients early access to start testing and giving us real feedback.

40:29 that's exactly what the open beta is for. And I want to encourage you to sign up for the weight list. if you don't yet have access to AlphaSense, I encourage you to sign up for a trial so you can get familiar with the platform before Super Analyst. thank you for your time today. I truly cannot wait for you all to try Super Analyst. And, I think we can now open it up for for questions. >> Sounds great. thank you both, Louis and Chris. This was incredible. And like I mentioned in the beginning of the call, we'll try to get to as many questions as possible. There are tons of questions in the chat. So, thank you all so much for submitting those. I will just start by asking, how hard is it to actually build a custom skill? Do users need to know how to code or write prompts in a very specific way to create these skills?

41:21 >> I I can take this one. Yeah, perfect. So, the it's very easy to create a custom skill. you can just ask in natural language. You can ask the super analyst, help me, help me build a skill you know to do earnings analysis or whatever the use cases that you have in mind. And then what happens when you when you type that in is the we have something that we're calling this kind of skill builder or skill creator. And essentially this understands the content, the metadata, the tools available on Alpha Sense and then will help you step by step guide through how to build a skill for your specific use case. so it's very easy. you just ask a natural language.

42:01 >> That sounds great. And maybe as a as a quick follow-up to that one, if a user has created a custom skill, can they share it with other team members? the so so that's a capability that you know we're aware of that plans are going to want to have and like that's currently not something that will be available in the beta but it's something that is on our road map for sure. >> Great. I'm moving on to some questions around formatting of reports.

42:27 How customizable is the overall formatting? can a user give it a formatted file and have it fill that out or does it always follow a specific format? Yeah, I can jump in. Definitely related to what Louie was saying, the Super Analyst is incredibly customizable. First of all, just taking a step back. The amount of questions coming into the chat was incredible. Thank you to everyone for the the the amount of engagement and and we definitely want to follow up with lots of you offline. But but yes, so so you can upload files to Super Analyst and build skills off of those files. So, so I want to upload an investment memo and say create a skill that's going to follow this particular format. a slide deck, a model, whatever, it is. Super analyst will read that file, understand the specifics of of the structure, the the aesthetics, all of that, and and and the content, the tone, language, and build a skill aligned to to your particular template.

43:32 >> Thank you, Chris. And I guess as a follow-up to that as well for any of the you know products that super analyst is able to create how are these being created? Are these you know models and Excel spreadsheets similar to what a you know human analyst would have created? Can you tell us a little bit more about how these initial templates are being developed? So yeah, I'll take the first step and then Lou, you you build off because because I want to talk a little take the opportunity to talk a little bit about how our organization is evolving. So you know as AI has become so core to AlphaSense everything we do you know we wanted to build super analyst as a power user of the platform and that meant hiring people who have really sat in your seats done the job and so we have really invested in bringing bringing in top tier talent from every one of our industry verticals who are the ones actually architecting the core capabilities of of super analysts to meet that sort of bar of quality that you have all learned to come to expect from AlphaSense. And so that's sort of the starting point. And then Louie, anything you want to add to that?

44:42 >> No, I think that yeah, I think that's a great answer. So like, you know, some of the stuff I think one thing that was pretty cool and I think you all saw this on the demo was also the fact that you know we have this corpus of of of canal built Excel models and you know using that as a base for some of the kind of financial modeling and some of the tasks. both in terms of you know in this case like starting out with that model to to build out kind of your own version and then another interesting example was you know we tested an example asking to build a you know private model for data bricks and you can then ask the super analyst you know use snowflake or the closest public comp model as a as a comp example and then kind of pull that in and use that as the base so there's also that existing kind of institutional knowledge and structure around excel which is pulled into super amist.

45:32 >> Great. Thank you. for some of these you know outputs we're creating, they are sort of they they can exist outside of the platform such as the Excel model once that is downloaded into a user's computer. Are users able to edit it directly and use the formulas included or would they then have to go back to super analyst and ask for those updates? Yeah, these are these are real deliverables. So, you know, the Excelss are are fully editable Excels. I'll plug one more capability, which is we are running a beta of our Excel and PowerPoint add-ins right now, which are essentially an extension of Super Analyst into those platforms. And so, the way we think about the workflow is, you know, a lot a lot of platforms out there show demos that sort of oneshot these capabilities. And of course, we think Super Analyst is going to be best-in-class at that because of the underlying content and the the the way we've optimized the platform to understand the the specific workflows.

46:33 But we know that's not the end. You're going to want to pull that deck into PowerPoint and iterate. And that's where those add-ins are really valuable. You can convert continue to converse with AlphaSense in natural language and edit those those decks and and models directly there natively. Great. along the same lines around outputs, you obviously showed some amazing things that Super is already creating. Can you talk a little bit more about what other products it'll be able to produce once it's ready? You know, what's what in the pipeline?

47:10 >> Yeah, for sure. So, yeah. So recap what we saw was you know kind of obviously obviously the in chat answers and then you know it opens up the canvas to show you the dashboard which is fully interactive and that dashboard you know you can configure and ask for charts and ask for different types of kind of visualizations on on that dashboard Excel of course we saw the Excel models and then in addition to that what we didn't demonstrate is you know the super analyst can create PowerPoint de so it can create slides I'll do a plug here for our slide slide agent. if you're using generative search, try the the new slide agent out and ask for for slides in your queries.

47:48 the superanist will integrate with that slide agent to create these really really polished slide decks and we're working on the ability for you to upload your own template and then have that super analyst essentially follow the instructions in your template. yeah, and then of course we saw the example where you ask for report, the ability to download that as PDF as well. We know that's important to be able to send that and share that examples with each other.

48:12 >> Great. >> Yeah, those dashboards today that you know they're what Louis showed you those are static dashboards. They're pulling data from from the platform. They're interactive but but you know they're not connected yet into the underlying data streams. But that's coming and that that we think is a total gamecher like we think of it as all of you are going to be Vive dashboarding in Alpha. you're going to be, you know, describing the the data and the transformations of that data that matter for you. You're going to be able to create an interactive dashboard in sort of whatever structure you want. and that will stay live connected to the the underlying data assets. So that's where we're going really quickly with all this.

48:54 >> Thank you. Yeah, I think digging a little bit deeper into that interconnectivity. We've had some questions around access to different content sources and how do we ensure you know the the agent isn't hallucinating. could you dig a little deeper into what exactly Super Analyst has access to on our platform for users who have maybe different entitlements and access to different sources? >> The Yeah, take a first stab there. the so essentially it super analyst adopts all of your entitlement permissions and has the exact same kind of entitlements that you have. So the it it'll essentially have access to the same content that you have and apply to the same restrictions around different content sources. again like I think the mental model here is like this is your power user and I think the framing here is like it's your power user with similar entitlements to what you have and it has access to the content and data that you can see and find on the alpha platform. and that's also a nice way to work with super analyst is if you know kind of where specific tools are and content sets live like you can navigate and use the super analyst to use those. yeah and then as I mentioned kind of at the end of the demo like we are excited about how to continue integrating these different tools into into the super analyst like what I showed now around watch list creation is is really exciting. and you can imagine like integrating more of those capabilities around being able to quickly generate a you know generative grid or add comments in the document or highlights or create a notebook and essentially take all of those same actions that you can on the Alpha Sense platform are things that we're thinking about for ser for roadmap items.

50:39 >> And then Sammy if I can just jump in on the hallucination part of the question I think that's an important one. So, so look, you you all do not want us to just create a a search summarization tool like like that's existed for a while. We we can do that. What you want is for this super analyst to start to make real inference on top of all the underlying content and that comes with with some risk. And so we've really thought about the the workflow of how you can trust super analyst and validate the outputs in a way that really fits natively into your workflow. And so, you know, Louie talked about the the the skills and rules built into Super Analyst. So, every fact that it's using, every assumption it makes is going to be listed and and cited into the underlying content. And it's not going to be click the citation and you load some website where you have to, you know, hope to to to pick around and find it. It's it's going to be cited in line just like, you know, the what you all expect with generative search. And then we're building specific skills to support the workflow. So we'll have an explain skill where you can just ask superanalyst to take a slide deck and reconstruct its thought process for how it found and sourced and constructed every slide.

51:48 That itself is is another kind of deliverable that allows you to quickly trust trust the outputs. in what Louie showed, he showed the Excel outputs, like those are separate artifacts in and of themselves, but you could like just like an investment banker would build a slide deck for a CFO and an underlying Excel workbook to support it. That's what Louie was showing. There's an Excel workbook that's that shows all of the underlying data assumptions all linked back to AlphaSense. we're also creating what we call a a critique skill where you can take any output of of AlphaSense and actually use a different LLM to poke holes in the argument and and try to to identify you know, weaknesses, opportunities to improve the argument.

52:31 And so that's an example where our multimodel approach really adds value. So we'll always be on the frontier with the best models and we'll be able to actually use the models to poke holes in each other's arguments. >> Great. Thank you. I think we probably have time for maybe one more question. I think there there have been some questions around uploading proprietary data, you know, their own content, maybe like local content. can we talk about whether that's possible, how that's going to be integrated into super analyst answers?

53:03 >> Yeah, I can jump in on that. So, so AlphaSense already allows you to integrate content and and we, you know, go through kind of infosc security requirements with with many of our, you know, clients that are working on sensitive proprietary work and and so we're able to do that in a way that gets them comfortable with it. And super analyst will will be the same. You can upload data to your superanalyst that's not going anywhere. It's isolated and and secured for for you to work on.

53:32 And then you ultimately have control of everything that lives there. So if you want to remove documents you know remove remove deliverables that's that's within your control as well. But everything is secure everything is isolated everything is owned by you and and nothing is used to train or or influence the broader platform that that you upload into the plat into into Super Analyst. >> Great. Thank you so much both. And again, thank you all for the many many questions you submitted. for anything that we didn't get to today, we're going to try to follow up separately. and I I saw that, you know, Wendy shared that link to access our weight list. So please sign up. that'll be available to our users whenever our beta opens.

54:17 in the meantime, you know, thank you all for joining us today. you will receive a recording of this webinar in about 24 hours along again with that link in case you missed it in the chat and if you're not a customer yet you will hopefully also get an outreach so that you can set up a trial account like Louis mentioned we already have all these amazing capabilities in officeense right now and we want you to encourage you to get comfortable with those and get familiar with those ahead of superanalyst release when this session ends you will also see a survey we would greatly appreciate it if you could take a moment to complete Your feedback is incredibly valuable to us as we continue bringing these types of webinars to you.

54:55 >> And one more Sam if I can add as you do get access to super analyst where we're going to roll out the beta shortly and and expand it quickly. Definitely raise your hand on getting on the phone with me with Louis with others on the team. We absolutely would love your feedback and for all of you to help shape where we go with Super Analyst. >> Yes. Thank you so much. >> Thanks a lot. >> Thanks everyone.