Good morning everyone. So I've been told that a full third of you are at your first AU. So can the other two-thirds welcome the third to AU?
And I have a question for the die-hard two-thirds. Does it feel good to be back in Vegas?
And I probably said this before, it was my well-intentioned but bad idea to rove around the country. So there you go. We listened. We're back in Vegas and we're here to stay. So welcome.
All right. The digital world is accelerating exponentially, but our ability to deliver in the real world, it just hasn't caught up yet. It remains constrained by money, by labor, and resources. And all of you are being asked to close that gap because intelligence alone doesn't build necessary housing and infrastructure, manufacture critical products, or create entertainment that inspires billions. People do. You do.
So, how can we help you accomplish more than what was possible before, but with the resources you have? Our industry clouds, Forma, Fusion, and Flow, expand what you're capable of and help you rise to the challenge. They're how all of you will change the entire ecosystem and how it works. They connect people, processes, and disciplines across the entire project life cycle. In other words, they help you close the gap between AI and physical outcomes faster than ever before.
Now, AI for design and make will only succeed if it's grounded in the realities of the built world. We know that what you need is project intelligence. A connected brain at the center that brings every data point, every decision, every discipline together. A brain that grows smarter as a project advances. That's why Autodesk AI and our industry clouds unlock your capacity across design, make, and operate. They work seamlessly with the tools you use today and they allow you to move incrementally to the tools you'll be using in the future.
Your loyalty funded all of that future work. And that's why if you're an Inventor and Revit subscriber, you always get access to Forma Infusion. We bring the future to you because you helped create it with us.
So today I want to show you what's in it for you, for your company, and for your entire ecosystem. Unlocking capacity starts with you, with a single designer, engineer, architect, construction PM, or VFX artist. We are eliminating grunt work with AI to make each and every one of you more productive, more creative, and more innovative. Because when you have the freedom to focus on higher value work, you can solve bigger problems and make a bigger impact.
So let's dive into design and manufacturing first. We are eliminating grunt work in Fusion. And I want to show you how AI is starting to touch every step in the workflow, supercharging the entire process. So imagine you're designing an autonomous robotic lawn mower. You'll see more of that tomorrow, by the way. And you want to bring in this front bumper. Files of existing parts often don't contain any information about how things were designed or constrained. But with auto-assemble and auto-timeline in Fusion, that's not a problem. Fusion will find the connection points and properly orient the assembly. It even generates an editable, intelligent timeline of the features required to edit the bumper. And it learns from parts in existing projects.
But if your new bumper doesn't work with the front assembly and you need to replace the bracket with a similar one that you've already designed, well, AI can help too. Find Similar in Fusion will identify and help swap alternatives, going beyond a simple part number search or basic geometry capture. It's AI powered by our industry foundation models. And once you're done designing, AI boosts your make workflows as well. Instead of machining this frame with endless picks and clicks, you will soon be able to use automated CAM programming in Fusion to create tool paths for you. And while you adjust its output, Fusion will pay attention, tailoring what it generates for you.
Now, eliminating grunt work like this isn't just possible in design and manufacturing. We're bringing AI productivity to architecture, engineering, and construction with Forma and Revit. Take for example these residential buildings. Let's say you have massing models and you want to generate architectural layouts from them. With the new building layout explorer in Forma, and this is underpinned by our industry foundation models, you can do that in minutes. It takes into consideration the project context and the unit mix requirements of the building to create an actionable result for you. With the layout in place, you will then be able to tap into AI-driven workflows like furnishing one room by referencing another one. And as your layout changes, Revit will let you remix room elements, keeping your projects up to date every step of the way.
And once you have a high-quality model of your building, documentation should be easy. That's why Revit will create and optimize annotations where and how you want them. They will even rearrange as your design evolves so that the documentation stays organized without the need for manual adjustment.
But we're not stopping there. To truly transform project delivery, AI productivity in Forma needs to extend beyond design. We're introducing agents for some of the most important construction workflows and powering them with design and make context. Our new document agent will be able to search across files and metadata right in Autodesk Assistant, helping you quickly find the insights you need and turning disconnected information into a connected source of truth.
Building on that foundation, the new submittal evaluation agent automatically checks submittals against critical specifications. It then provides clear compliance recommendations so you can review, override, and decide before you approve. And that same project intelligence extends upstream with the bid proposal agent. It extracts data from subcontractor proposals and then maps it to the scope of work so you can compare proposals more easily and make informed decisions with confidence.
Now I think we know there is one industry where AI is more prevalent than any other and that's media and entertainment. So we're adding new AI features more rapidly to accelerate your creativity. With Flow, our industry cloud for media and entertainment, you can generate characters and entire worlds. You can now use Wonder 3D and Flow Studio to input a simple text prompt or image and then generate a character like this one.
You then bring your character into our new Flow Studio 3D editor, automatically animate it based on a real actor's performance, and then apply a camera move. This workflow alone usually takes artists weeks. With Flow, it can be done in just a few minutes.
But a character isn't complete if it doesn't interact with a scene or a world. That's where our partnership with World Labs comes in. World Labs is building spatial intelligence through multimodal models that perceive, generate, and interact with virtual and physical worlds. And our teams are working with them to bring their model called Marble into our tools. With Marble and Flow, you can again use a text prompt or image and an entire world gets generated for you. Now you have a camera, a character, a camera move, and a world all in our 3D editor where you can orchestrate them.
And this workflow extends to expert tools as well. With just a few clicks, you can bring your generated world into Maya, integrate it with your existing shot, even one with crowds, and get the exact creative output you were envisioning.
And I think you probably recognize technology like this isn't just for VFX artists. For example, architecture, engineering, construction teams can reconstruct a real site from just a handful of photos. And if you want to learn more about Marble for AEC workflows, check out the class on Wednesday from World Labs co-founder Ben Mildenhall. Because regardless of the industry you're in, these tools let you iterate and explore ideas while staying in control of the creative process. They make you more productive and more creative, allowing you to focus on more impactful work.
Now, unlocking capacity doesn't just benefit individuals. It scales to teams and entire companies. It makes outcomes more predictable and helps make companies more profitable and competitive. Because in a world where everyone faces the same constraints, capacity becomes the advantage that separates leaders from everyone else. That is where our industry clouds' ability to connect data and context comes into play.
Last year, we announced Forma as the first comprehensive AEC industry cloud. By bringing Autodesk Construction Cloud into Forma and creating a single project ecosystem, we're unlocking capacity for companies across design and construction. Since then, we've been working tirelessly on more connected workflows. Forma for buildings, our complete solution from site to detailed design, is becoming more powerful.
To show you, let me yield the stage to our live demo presenters. And yes, we're doing more live demos again this year to kind of show you how the products really work. So, please welcome our Forma experts, Robin, Yasper, and John.
Thanks, Andrew. I'm back and better than ever. So, today we're going to be working on a master plan site in Littleton, Colorado. It's a mixed-use development with a hospital, residential, and commercial uses. As an urban planner, I know that the earliest decisions on the site are often the most important and the most expensive to get wrong. Yet, we usually make them with the least amount of data possible. With the new site design experience, we're able to get contextual data right from the start. And new this year, we're able to pull in project-specific data directly from Forma data management into site design like this DWG. With the layers intact, there's nothing to convert and nothing to clean up. Also notice in the new site design experience, we can toggle between 2D and 3D seamlessly.
Okay, so I've already populated some buildings here using the explore tool and added relevant functions to get a good overview of our gross floor area, rentable area, and floor-to-area ratio. I'm going to focus on these two masses on the northwest corner of the site. All right, watch this. I'm going to activate the road tool and boom, the masses split from two to four. That's new this year, reactive modeling. Pretty cool, right? I also want to make sure that these masses are following the relevant constraints. So, I'm going to go ahead and add right-of-way as well as a slope setback to ensure that we get adequate daylight on the ground floor uses.
Zooming out, we can review the site as a whole. In addition to reviewing the density metrics, I can also ensure that my parking demand has been met as well as our land use distribution. And of course, I can run a daylight analysis. One click tells me which buildings are performing and which aren't. And luckily, I have an architect here to dig into one of the apartment buildings. Over to you, Yasper.
Thank you, Robin. So, I'm in the same project, but I am in for a building design. What Robin just handed me is an empty shell, a massing model, right? And as an architect myself, I want to start digging into how to populate this building. I am going to go and look for my sort of known units, my known apartments. I start there and that's new this year. You can add a library of your preferred apartments to Forma building design and use those to design with. I'm also going to keep an eye on that daylight, Robin. Now, I've cheated a bit. I've added an array of different kinds of apartments. Some small, some large, some medium. And I can drag and drop these out. Start filling out my building. I can add cores, of course, and corridors. I'm going to grab some of the medium and smaller apartments and add them to the west side. And as you can see, it doesn't quite fit the building, but that's easy enough. I just adjust the size. Another corridor, and we're on the home stretch. Facing east, I'll do some of the smaller apartments again, pulling them out into what I would like for an array to be, adjusting the facade to fit. And in the south facing the sun, I'll do some of the more larger, more premium apartments.
Of course, I'm going to have to add some cores. Now, where I'm from, we get away with one set of cores, but we're in the US, so I'm going to grab this one. I'm going to grab a core and flip it. So, we have dual point of view. As you can see, some of these don't quite fit. So, I can drag and adjust on the spot to make them fit the way I want. Now, let's have a look at the interior daylight, which is new this year. We're running interior daylight where we measure lux from every window in every room as we design. This thing allows me to check and adjust my design as I go instead of checking it afterwards. If you're looking for LEED points, WELL points, this is a very good place to start. Now, we went from an empty shell to a full building, but 2D is one thing. We can of course also look at the same building in 3D. Now, as you can see, I can now look and browse through every floor of my building and see daylight and lights as we go.
Now, we've got our facades, we've got our balconies. I can dig into my roommates. I can look at everything down to how many bedrooms I have or whatever else I need. Now, for me, this is as far as I'll go and I would continue in Revit. However, I have this retention pond to the north of my building, which I think I need a civil engineer to have a look at. John, could you help me out?
Yes, for my friend. It is a detention pond, not a retention pond. It's going to be a dry pond, man. Sorry, man.
I'm the civil engineer on this project, and I need to take a look at how the proposed buildings are going to affect the existing drainage system. Like Revit, Civil 3D is now a Forma connected client. That's right. I said it. We're all together again and we're all connected. You know what that means for architects, engineers, and planners? That means that we can all take advantage of that great coordinated project data in your environment, in your applications. I can see all of that directly inside of Civil 3D. So, we need to fix this problem. We need to make this detention pond a little bit bigger because they have increased the impervious area. So, we need a bigger pond. So, let's do that real quick. We're going to go in here and use a feature called creating a daylight feature line. We're going to grab our detention pond and we're going to add a bottom. We're just going to make it bigger. All right. 8 feet should be enough. That's a pretty deep pond. We'll go ahead and add that particular feature line to our surface.
All right. And our pond is built. Now, I want to share this with both Yasper and Robin. All right. So, I need to push this back to Forma. All right. So, I'll do that real quick. Now that it's been pushed back to Forma, they should be able to see this information in their model. Let's see what happened.
Yep. Let's check. Yep, there it is. There's the pond from Civil John. I dig it. What would we do without you? As you saw, Forma kept us connected to the latest data. It doesn't matter whether you work in Forma for buildings or a connected client. Everyone moves faster together. Back to you, Andrew.
All right, let's give a hand for the demo team. So what did you just see there? All right, what you saw there was Forma acting as a single database connecting design all the way through construction, starting with BIM but ending with project intelligence about what was going on. It doesn't do everything Revit does today and it won't for a while, but it connects to the tools you use every day. Last year we showed it connecting to Revit and this year I'm thrilled that Forma is expanding into the civil infrastructure space. It brings all these different disciplines together, letting them iterate together from the concept all the way to BIM on into construction. It's a great example of how Forma can provide context across the full life cycle of a project, making outcomes more predictable and companies more profitable and competitive.
Now, let's take a look at a similar scenario, but in a different industry. We continue to push the boundaries of Fusion, our industry cloud for design and manufacturing. Like Forma, it brings together data, people, and processes in the cloud. And now we're unlocking capacity for companies by reimagining PLM, otherwise known as product lifecycle management. And we're doing it by helping manufacturers get past the billions of dollars they lose on PLM every year because of slow approvals and clunky tools. I'm excited to introduce a brand new Agentic PLM experience in Fusion that provides early insights and accelerates decision-making. So, let's see this one live. Please welcome our PLM experts, Jeremy and Jennifer.
Thank you, Andrew. I'm super excited because I get to be the one to share with you and show you the new Agentic PLM experience today. It's going to make PLM a lot more powerful. It's going to make it easier to use, and it's going to make it more valuable for everybody within the Fusion ecosystem, regardless if they're a CAD user or not. So, for today, I'm playing the role of a supply chain manager, and I just received an email from our vendor, and it's telling me that we're going to be experiencing a shortage of motors, right? And that's bad news for us because it's going to delay our production schedule. Now, what's good is that I don't have to rely on old school clunky PLM solutions anymore. Instead, all I need to do is bring the email into my new re-imagined PLM experience and the assistant is going to extract all sorts of information that I can act on. I'm also really glad that Jennifer is here with me. Jennifer is my engineering manager and Jennifer is going to help me with all the technical issues and details that I need to solve this problem. Good morning, Jennifer.
Hey, Jeremy. Jennifer, I'm going to invite you to this workspace so that you and I can work on this problem together. This new Canvas experience is a fantastic place for teams to be able to work together to solve problems. I'll be back over to you in just a moment, Jennifer.
Sounds good. Ready when you are.
Okay, so real quick, let's take a look at what the assistant has provided me with regards to information. I can see here I've got my Fusion data available to me. I can see that the assembly here that's potentially impacted is available. I've got the email which gives me more context. I even have the spec sheet of the motor that's being affected. And if I look here, I can see that the assistant has provided me with some impact analysis. So, this is really going to help me understand what's going on. I can see a single motor is affected and only one assembly has been impacted. So, that's great news. But down here, I can see, oh man, there's 2,400 motors that we're going to be short. So that's bad news. And to make matters worse, we don't have any approved alternatives either. Okay, so we need to figure out what we're going to do to be able to solve this problem. Now, this would take me a long time to figure out going through a variety of applications and data sources, but with the assistant, it makes things much, much easier to be able to resolve. All I need to do is ask it to find me some substitutes that still meet the product's requirements. Now, this is pretty technical, so I'm going to ask Jennifer for the engineering perspective. Jennifer.
Awesome. Well, let's take a look at my screen. Now, in the past, when Jeremy asked for my engineering expertise, I had to go hunt down information from lots of different systems, even to understand what he was asking. But with this new Agentic PLM experience, the issue, the analysis, and some of his recommendations are all laid out in front of me, making it easy for me to get up to speed. Okay. So, in order to start to make a decision, I want to ask the assistant to give me an overview of how this change is going to affect our bill of materials. Our bill of materials is a list of everything that goes into our lawn mower. And at the end of the day, if something isn't accurate in there, we risk production getting off track. Okay, so we've got this bill of materials. Now, it's time to start thinking about substitutes. Hey, Jeremy, where are we at with those?
Engineering teams are always so impatient, aren't they? All right. Well, the good news is this used to really stress me out when you guys push me. But with the assistant, it's made things much, much easier. It's already given me three options to consider. The second option looks like it has the best lead time, but the third option has the lowest supply chain risk. So, I'm just going to document that one and we're going to have the team move forward. I think we should be good.
Hey, now I think that's an engineering call. Why don't I take a look at this one? Good point. Okay. So, let's take a look at this bill of materials. Now, two important things here. First of all, I stay in control. I can mark up and refine this bill of materials based on Jeremy's recommendation. And second of all, I have full traceability. All these decisions that we're making are being logged from the start. But honestly, what I can't get over is the fact that all I did was ask the assistant to update and it automatically added in these bolts, nuts, and washers. Do you know how much time that just saved me? Okay, before we wrap up, I want to ask the assistant to draft a change order and assign tasks so I don't have to individually email every engineer that's going to help us make this change. Now, this used to be a series of tedious administrative tasks that's now automated for me, which really is saving me a ton of time. You know, supply chain problems used to be difficult for us. Honestly, they used to scare me, but with this new Agentic PLM experience, I feel like I have a partner that helps me move forward with confidence. Back to you, Andrew.
All right. Thank you. Another hand for our demo presenters. Now, I've never met an engineer that says they love their PLM system. Really, you know, the PLM system kind of gets in the flow of what they're trying to do, but it really does provide important value. And since Fusion, just like Forma, is one connected database across design and make in the cloud, we can layer on powerful agentic experiences that make it really easy to do hard things and get the value without the complexity. Our new solution in Fusion makes PLM so simple, it even makes it invisible. It unlocks capacity to collaborate and solve problems across teams. Again, this makes outcomes more predictable, increasing profit and competitiveness.
Next, I would like to dive a bit deeper into agentic experiences. They're fast becoming the backbone for collaboration everywhere. Now, let's first hear from some of you about how AI is changing roles and industries and then invite our CTO Raji Arasu on stage to show you how we're responding with Autodesk Assistant.
Wonderful to be back here with all of you. Our teams at Autodesk have one mission: to unlock capacity for you. For more than four decades, we have helped you design, make, and deliver with precision, confidence, and accuracy that the physical world demands. But probably right is still wrong. And in addition to this industry expertise, our investments in areas like cloud, AI research, machine learning, and more have created a strong foundation that sets Autodesk AI apart. This expertise helps us derive project intelligence, understanding not just what's happening on your project, but why and what to do next. Intelligence that generic AI does not have.
Autodesk Assistant is built on that foundation and that intelligence is designed to understand your work and work alongside you in the company workflows where important decisions are made, which gives you more predictable outcomes, more profit, and more differentiation. Last year at AU, we told you that Assistant would be a powerful teammate working by your side with you and your teams. And we delivered Assistant in all the major projects you use every day.
Since March, you and we have seen 2 million conversations and a 99.9% success rate. Launching Assistant was just the beginning, the start of a learning loop. Every week we listen, we learn, we improve, and every week we keep raising the bar on quality and experience. Across industries, customers are telling us they're seeing real gains, cutting design time in half, spending less time searching and more time solving the actual problem. You saw it yourself just a few minutes ago in the PLM demo. Assistant is now finding solutions for you instantly and that's real impact.
But we know at your company level important questions still remain. Let me share with you the top three questions that we have heard from you. Are the answers accurate, actionable, and grounded in your data? How can you connect your company's AI with our Assistant? And how can we scale AI together, keep it fast, accurate, and most importantly, cost-effective?
Let me start with the first question. To answer that, I'm excited to introduce Autodesk Context, a capability that makes AI more actionable and accurate. It's not pulled from a prompt or a siloed data source. It is built from a deep understanding of the workflows, projects, and the thousands of decisions your company makes every day. And it's not just about tracking what happened. It is why it happened. And that's what makes Context powerful.
You have all lived this. A project finishes and then six months later, someone asks, 'In design and manufacturing, why did we swap that part? In AEC, why did we move the wall? Or in media and entertainment, why did we cut that movie scene?' The system may know what has changed. But why was that decision made? And too often that context is gone. Locked in someone's head or it's scattered in some system across the board. Autodesk can help solve that. And now in addition to the why, we are also capturing the when. When is not just a timestamp but the state of the workflow at the moment when that decision was made. The exact situation, the constraints, the data points, the people that led to making that decision. These two pieces of information make Autodesk Context really powerful for our industries.
Andrew described a connected brain at the center of your project, growing smarter as a project advances. This is that brain. It is a map of how information, people, decisions, and activities in a project relate to one another. And all our work on granular data models power this context. Every element, every relationship, and every project record. When Assistant understands this context, the agentic experience that we have built on top of this layer stops being generic. It becomes grounded in your workflow, your company, and your project ecosystem. Let us show you an example. Working with customers like DPR, Autodesk Context is making Autodesk Assistant more accurate, more relevant, and more actionable than a generic LLM. Please welcome Autodesk construction expert Lori to demonstrate.
Thank you, Raji. For this demo, I am in the role of a project manager on DPR's Silicon Valley Office Build. But if I can be honest, my title might be PM, but what I actually do is worry for a living. I obsess over every way a project can go wrong and fix it before it happens. All to keep this project on finished on schedule and on budget. Simple, right? Except, as we know, schedules don't usually blow up over one big disaster. It's a one-line email combined with an RFI. Small things that add up to be big things over time. And when this happens, I need to know two things. How is this going to affect my project today? And what's going to happen next week once it's too late to fix? And that's not really a go look it up problem. If I'm going to use AI on this project, it needs to know my project, the elements, the activities from my model and my schedule. And in any case, if I'm going to walk into that meeting with the architect and the owner, I can't really say the AI said so. I need evidence. And so I turn to Assistant and I ask one simple question.
What is the impact of RFI 169? Within seconds, Assistant begins tracing that RFI across the project. It identifies the affected building elements, the related schedule activities. And this is so much more useful than just having a list of things. And look at this. We can see the shaft wall enclosure at the main electrical room. There's no approved fireproofing detail where it meets the exterior facade. And would you know that this facade is going up this week, which means delays, rework, reinspection, money, and time. And this is amazing. I can look at this. I can see it. I can plan. But that's not the really cool part because Assistant continues to trace the ripple through the project. And that is what helps me sleep at night. I can look at this and see and explore impact chains. I can see when level two facade might be affected. I can review level three. Am I going to have to leave a whole section of panels out? What's that temporary weatherproofing going to look like? I can inspect the main electrical room and start understanding. Can I protect that area? And we can see how one RFI might ripple all the way through to power the project.
And I can go further. I can ask follow-up questions and really understand each risk and each mitigation. And this is the information I need, the real data from my project. This is what gets me ahead of the problem and prevents that change order or that delay. And that is the difference between AI that can talk about construction and an Assistant that truly understands my project. It turns 'we'll find out' into 'we already know' and that is cool. Back to you, Raji.
Thank you, Lori. I'm so glad Assistant is helping her sleep better. And that is the difference Context makes. Not about sleep, but a generic language model can tell you what to check. Autodesk Assistant uses Context to trace what changed, what it affects, who needs to act, and why. This means less searching, less backtracking, and more confidence in the next action.
But real work doesn't live in one system. Your company has more context that lives outside Autodesk. Context that lives with your teams, your tools, and the partners you already trust. And many of you are already building AI on that context. So the next question is how can your company's AI work with Autodesk Assistant? Autodesk has always believed in open solutions that help extend your unique workflows. Agentic is no different. So today I am introducing Autodesk Assistant Builder. It is designed to let you plug your agents and your partners' agents into Autodesk Assistant. Your AI, your partners' AI, and Autodesk AI all working together inside Assistant, giving your team one powerful experience. So, let's see that in action with an example from Arkil. That's one of Denmark's largest engineering firms working on buildings, infrastructure, and beyond. Arkil's team is designing a college dorm in Copenhagen. Their structural engineer has just completed their first review and it looks like they have found a problem. To help solve this issue, an Arkil admin will use Autodesk Assistant to add Arkil and its partners' AI connectors right into the Assistant making that experience seamless. And for this demo, let's welcome Alex and see how Assistant can get him and Arkil out of a tough situation.
Thanks, Raji. For today, I'll be the structural engineer at Arkil. And there is so much on my plate right now. Our first review just came in for that dorm project, and we're already over our embodied carbon threshold. That's not good news. That could add weeks to the project. I need to pull up carbon analysis, do cost calculations, coordinate with our project team. Instead, how about we get the sustainability and the engineering right in minutes, all without leaving Revit. Let's give it a try. I open up Autodesk Assistant in Revit and see that thanks to Assistant Builder, there are two new connectors available. Arkil, our own project data, and also NTI, a third-party sustainability platform. Not only that, I see that Assistant has three suggested prompts personalized to use with these connectors. This top one sounds great. Suggests three ways to reduce embodied carbon. Assistant gets right to work. It's combining the model in Revit with NTI's sustainability platform to help me cut embodied carbon. And instead of just pointing out the biggest carbon source, it identifies one that I can actually isolate from the floor system, so I may be able to swap them without forcing a larger redesign. Let's learn more about those columns.
Once again, Assistant looks at the Revit model, looks at all the structural steel columns, and finds that there are 21 identical columns that we can evaluate together. Let's find alternate material options for this column type. Once again, Assistant is connecting to NTI to help me cut embodied carbon. It looks like concrete wins on carbon, but the timber material is tempting, too. It's lighter and faster to build, so it's worth considering later when I do my structural checks. But of course, changing materials affects more than just carbon. So, let's pull cost into this as well. Now, Assistant is connecting to our own project data, what we've actually paid for materials and installation on past projects to help me estimate the cost of these changes. Just a few clicks and I got what would have taken me a week of back and forth. I still need to do my own structural checks, buckling, connections, and fire rating. And I need to coordinate with the project team. So I ask Assistant to summarize the conversation and I can send the trade-offs to the project team.
We decide that the timber material finds the right balance between structural and carbon and cost. So we want to move forward with that. I can come back to the same conversation in Assistant and make it official. Steel to timber. I like that Assistant didn't make this decision for us. The engineering decision and the cost trade-off was up to the project team. Thanks to Assistant Builder, we went from what could have taken weeks to actionable results in minutes. Back to you, Raji.
Thank you, Alex. Notice what actually solved the problem. NTI knows carbon. Arkil knows its own project history. And then Autodesk knows the model. None of us could have answered Alex's question alone. He never had to leave Revit to get there. What took weeks across multiple systems just happened in one connected workflow. That is the power of an open Assistant with Autodesk Assistant Builder. The more intelligence you connect, the more capable Assistant becomes for your teams.
Now the third question, one the entire industry is wrestling with. How do we scale AI in a way that's fast, accurate, and cost-effective? The model landscape is changing every week. Seems like every day some days. New options, new capabilities, and often higher cost. And no single model is best for every job. Your work is not just words. It's geometry, its schedules, materials, simulation, and so much more. That work needs specialized models optimized for speed, precision, cost, and deep domain understanding. I am certain the agentic future isn't one model. It is going to be a governed set of multiple models matched for the right task at the right time and at the right cost. Your teams should not have to evaluate every model or benchmark every workflow. That is our job. So today we're bringing you another capability called Autodesk AI Orchestrator. Our answer to your third question. Behind the scenes, Autodesk AI Orchestrator constantly tests models. It includes ones we build ourselves for engineering and design like I shared with you many years ago like neural CAD and more frontier models. It's always checking these models for accuracy, for speed, and for cost. And then it routes your work to the right intelligence. Sometimes one model, sometimes several, always grounded in the workflow. This is possible because it is built on our project expertise working with you.
to create those more efficient workflows. So for you, the experience stays simple. You get the right intelligence for your workflow while we do the heavy lifting. And that is how AI scales. Practical, trusted, cost-effective. What you just saw here is more than a smarter assistant. It is a new way for AI to work across design and make and it answers the three questions we started with. Can my company trust it? With Autodesk context, AI builds trust by deeply understanding your work. Can my company connect to it? With Autodesk Assistant Builder, your AI can work with ours. And can my company scale it? And with Autodesk AI Orchestrator, we route it to the right AI model for the right workflow. So you can scale in a cost-effective manner. With Autodesk, the answer to all of those three questions is a resounding yes.
And that is what we're building with Autodesk Assistant. Less searching, less stitching, less AI complexity, more creating, more solving, and more innovating. That is assistant today. Now, just wait until you see where assistant goes next. To show you, please welcome Andrew back to the stage.
Thank you, Audrey. So, if there's any idea that you take away from what you just saw, I want you to take away this. There is real technology, real expertise both in the domain and in the technological capabilities that is underpinning our pursuit of bringing you project intelligence that knows why you're doing things, what you need to do next, what you've been doing, and how you might act moving forward. And we're trying to do this all in a fast, accurate, and cost-effective way. And that's why I'm really excited to now talk about the experiences we're going to bring you in the future. And that experience is going to be through this preview of what we want to show you as the next generation of Autodesk Assistant. In the not-so-distant future, Autodesk Assistant will be available as a standalone agentic-first experience, one that reaches across products. And this is a really big step, but I don't want any of you to worry. We are not going to replace our in-product assistant or any of the tools and interfaces you rely on today. The next generation Autodesk Assistant will be a new option to explore, a new way to work when and where it makes sense for you. And we want you to help shape what it becomes.
Because I believe the greatest potential of agentic experiences is their ability to reach across products and even organizations. They can unlock capacity across entire project ecosystems so that everyone can commit fewer resources per project, bid on more projects, and win with greater confidence. Now, this is truly going to change how you work across disciplines and project ecosystems. So, I really want to give you another live demo today and a preview and I hope you're as excited about this as I am and you get the sense of how this removes complexity from what you do every day. Please give a round of applause for our demo team, Michelangelo, Tom, and Jeremy.
Thanks, Andrew. I'm so excited to show you this new next generation Autodesk Assistant. Now, before I get started, I want to give a quick thank you to Ajon Romeo Architecture for allowing us to use one of their really cool projects for our really cool demo today. Now, for the purposes of today, I'll be playing a totally calm, not stressed project manager working with this small but mighty team of Tom and Jeremy. Let's go.
All right, let's do this.
Thanks, guys. On mixed-use projects. Now, as a project manager, I wake up every morning feeling already behind. I have hundreds of emails and RFIs and issues and messages, and I have to figure out which one to address first. But not anymore. Autodesk Assistant has found the most critical items from across my projects, intelligently assembled by agents just for me. Here it's flagged a critical change from a client on a residential project. Let's open this in an assistant space. Now spaces allow me to curate data from across projects and products. Bring agents to create artifacts and do deep focused work. Let's see what this change is all about. Okay, it looks like the client wants to increase the rentable area by moving the south wall out 3 feet. Yeah, this is going to have some serious implications, right? I know we can ask the assistant, but instead of typing, I'm just going to talk to it. How will this change impact the schedule and what are the downstream impacts? Now what's exciting here is assistant is analyzing the underlying context then creating an artifact for me that will show me the impact on the schedule as well as the design. Let's see what it creates for us here. Okay, great Gantt chart with a lot of red dots in it, a lot of hot spots which tell me we may have some trouble delivering those interior layouts on time. I know what to do. Assign this issue to the architect. Now, I know if we can get this started on time, we may be able to get ahead of it with help from assistant. Hey Tom, you mind jumping on this for me really quick?
Uh, sure. Looking at that Gantt chart, I can kind of sense the stress coming off you, but you're not wrong. This is going to be tight, but with assistant, we're going to be able to get this done. So, we want to look at this change in overall area and what impact it's going to have on our room layouts. So, we'll go to a 3D view to get a better look. And I can select one of these units as a test. And I can ask assistant generate an LOD 300 apartment with walls and doors. All right. So, it's going to engage with our industry foundation model that we've trained specifically to understand and generate building geometry. And as it streams in the results, we'll see it update in real time in the 3D view. All right. All right. So, this is a kind of a good start. Now I can move on to the next task and I can just ask assistant generate apartment layouts for the rest of the building. All right. So now assistant will understand the overall context of this building and it'll be able to fit in these room layouts into all these differently shaped and sized units. Okay, great. Pretty nice, right? So, this is the kind of thing that would have taken me all day before, but with assistant, it got me to a really good starting point in my design in just a few minutes. Thanks to Autodesk's AEC industry foundation model. All right. Well, this saved me so much time. We can move on to another request from the client. So, guys, this will be fun. They're asking for this like slatted facade design. So, let's look at the sketch. Okay, interesting. So, here's the fun part. I get to use this handy agentic workflow that I created with assistant builder. It's going to help me with this design. So I can ask assistant run the slatted facade workflow using dimensions from this attached image. All right. So it'll look at this image. It'll extract any dimensions that it finds and apply them as inputs to this workflow that I had defined earlier. So, it found that projection dimension and give me like a UI that I can adjust here. Great. Now, it also has the slat density. I'll lower that a little bit. I'll run the workflow. And now, assistant will create some artifacts for me. So, I can inspect the design solution that it's generated for me. So, let's take another look here. Okay. Interesting.
So, it looks like I've got some brackets here that are going to be required for this more complex but better looking facade. So, you know, this is really powerful. I was able to define that workflow once and then assistant handled all of the deterministic steps and the agents for me behind the scenes. Super cool. All right. So, now I'm looking at some of these brackets look like they may need to be custom designed, but that's why we have Jeremy here, our design engineer. So Jeremy, I'm going to need some help designing some of these brackets. So do you have all the images and notes on your end?
Well, let me see here in my space. Looks like the artifacts have come through.
All right, great. It's all in your hands. Don't let me down.
I'll do my best. So, as we can see, Tom has sent me some specifications along with an image of the bracket they'd like to use. So, I'm going to use assistant to find a similar bracket that I've already built. It goes in, looks in my data, and examines all my different fusion designs and gives me options so I don't have to start from scratch. There's some great options here, but let's go ahead and select this 90-degree bracket. And you can see it pops up and I can look at even examine all the dimensions that are currently editable. But let's use the assistant to review the requirements Tom sent and apply them. So I'll just prompt the assistant to say, 'Hey, can you take a look at that document again, find the proper parameters and apply them?' So you can see it move and update. The assistant goes in, applies those dimensions, and the model updates automatically using Fusion as the backend. So, it understands the context of the model and how we want it built.
Really cool. Okay. So, now we've got this. We're going to make that configuration. The configuration happened automatically. But Tom being Tom, he wants perfection.
Or this new facade. Yeah. We got to get there for the customer. So, he sent over an image of some possible fascia to cover up some of the exposed structure. So, let's go back to the assistant again and say, 'Hey, can you generate some 3D models based on these images?' And it goes in, uses our industry foundation models to generate some examples that we could use for those fascias. And it's coming through. There's one, two, three, four different options. And it gives me some great visualization of those different pieces. Well, we got A, B, C, and D. We'll go ahead. Always pick a guys. What do you think?
Some cool fillets in there with variable radiuses, a little bit of angles. Looking pretty good. Okay, so this is what we want to use. And I've got an actionable model to start prototyping from. But before I do, I better communicate back to my teammates and say, 'Hey, let's give the assistant another prompt here and say, let's package up these results.' So assistant goes in, finds everything that we did and packages it up. And because this was a native fusion design, you see the artifacts generate automatically. We've got the design for the fascia and that bracket with its drawing ready to go. Guys, what do you think of the designs?
You know, I don't say this very often, Jeremy, but perfect. Looks perfect.
Yeah, that is so awesome. Thanks, Jeremy. It's so cool to see what can happen when engineers and architects can work so seamlessly together. The next generation Autodesk Assistant assembled project data and agents. It helped us reduce risk and it sped up everyone's work across products and industries. This is the future of design and make and we are so excited to shape it with you. Back to you Andrew.
All right. Thank you demo team. So what did you see there? Right? What you saw were people focusing on problems, things they were trying to solve, not trying to find out the capabilities they needed to solve them. This environment actually brought the UI, the tools, the capabilities to the user. Some of those tools were microservices that are agent accessible that can be called and run in lightweight ways. So the user focuses on the outcomes and not the tools. More capacity through more productivity. Pretty exciting stuff. And I think we can all imagine the amount of resources per project an agentic experience like this can save. One conversation, minutes, not weeks. The same small design and make team. Of course, why stop at make? One of the greatest sources of value isn't just in designing and making assets. It's how they're operated over years and often decades. The more life cycle phases we can converge, the more capacity we can unlock on your behalf. That's why we're expanding our investment in operations. Owners are a critical part of that story. When they're involved earlier, better decisions get made and more knowledge carries forward because the people who ultimately operate the asset can help shape it from the start. Their priorities, requirements, and real world experience become part of the design and delivery process, not an afterthought. More knowledge carries forward, fewer insights are lost between phases, and teams can make smarter choices with the full life cycle in mind. That creates continuity from design to construction to manufacturing to operations and ultimately better outcomes. To see what that means in practice, let's welcome our operations experts, Richard, Sarah, and Tim for our final live demo.
Thanks, Andrew. Okay, let's go back to our lawn mower project that Jeremy and Jennifer were working on and show you what our operation solutions can do here. I'm the production manager on this project. So, I'm responsible for actually making the LM500, which is the lawnmower model that got impacted by the supply chain issue. That change adds a whole new machining step to our production line. And on top of that, people are ordering the LM500 like crazy. These things are flying off the shelves. Our order volume has gone up by 20%. So, we need to ramp up our production from 50 to 60 lawnmowers an hour. And that is no small feat. I'm going to need to run some analysis to make sure we can get this done at cost. I have my A team here with me. Richard is our industrial engineer.
Hey Richard. And Tim here is our twin manager.
Hey Tim. Tim is new to the team. Tim, I know you've been working on building out our new digital twin. Is that something that could be useful here?
I thought you'd never ask. Yeah, absolutely. So when I started on the digital twin, I focused in on the advanced manufacturing center because that's a new building and we have the models and clean data. But the rest of our campus is over 30 years old. And the data is pretty messy.
Yeah. But that's not a problem now because Tandem supports the data that we have and we can start to build the digital twin directly from that. So, I've incorporated the floor plans already, and that's started to break down our location structure and set things up for us. And now, I'm going to go ahead and get the assets in for the fabrication line that you have that impact on. So, I'll go ahead and grab our spreadsheet of data. And this kicks off a brand new onboarding experience where before I used to have to do a lot of data mapping and it was really tedious. But now I can just ask assistant to help and it understands the context of all of our information in the digital twin and it knows how to map it in. So it's going to make some recommendations. Cool. While that runs, let me show you around a bit. Okay. So the other really cool thing about incorporating our floor plans is it actually builds out a spatial model that we can start to structure our data against. So all of our information that comes in, we have an understanding of the location breakdown, the systems and the asset classification, that sort of structure. So with that we can then anchor in all of our documentation like O&M manuals, our connections to machine sensors and IoT devices as well as our tickets so we have full maintenance history. On top of all of that information, Tandem now lets us build out our business metrics so we can focus in on what matters most.
Yeah. So everything understands how it's connected together in the digital twin. And assistant works within this context to help surface that information. So speaking of, let's check out our results here. Okay, so assistant has worked through our classification and applied categorization. It's mapped in our properties. It's aligned with the locations that were built out from our floor plans. And I think everything looks good. So I'll go ahead and import. All right, Sarah. I think the twin's set up for you. And you can explore this information in 3D, 2D, or even 1D. Or you can surface any of this in a dashboard.
Well, Tim, you had me at dashboards. The first thing I'm going to do is ask assistant to create a production readiness dashboard for the line that needs to change. Okay, while that's running, I want to go ahead and explore all that data that you just imported. So, I can see here that I have all of the data about my factory organized by location, which is super convenient. I'm going to click into the fab shop and then into the cutting and forming zone, which is the line that needs changing. And let's see what I have here. I have all of my machines that are on the production line.
Maybe load the streams panel.
Okay, let's look at the streams here.
All right, let's try refreshing.
Click refresh. Sometimes that helps. Yeah, that looks good. Okay.
Let's try doing this one more time.
Yeah, one more time. Give it a shot.
Uh, okay. So, what else could I explore while we have this open?
