AU Focus: Autodesk Platform Services and the Future of AI

Share this Article

Advances in artificial intelligence (AI) and machine learning (ML) have recently brought the potential of these technologies into the public eye. Why now? It’s in part because companies today have access to massive amounts of data—and the computing power to analyze it—thanks to cloud-based platforms.  

Autodesk researchers have been exploring these possibilities for years. As a result, Autodesk Platform Services (APS) gives teams access to a broad file format, from which data can be extracted for a variety of purposes. And APS connects design models with outside tools so developers can build new applications like digital twins driven by AI, ML, and other technologies. 

Here are just a few examples of applications that teams in Autodesk’s developer ecosystem are building on top of APS. 

An AI-powered design assistant  

Codeo is on its way to building an AI-powered design assistant for Autodesk software. Think of something like Siri for Inventor. Here’s the back story on how these Turkey-based programmers are getting the work done. 

Codeo’s customers design and sell highly customized, made-to-order products like playground equipment and kiosks. The design process, by nature, is highly iterative and time-consuming. 

Customers are always looking to move faster from design to product delivery. And speed is what Codeo’s voice- and text-powered app provides. 

Using a combination of the APS Design Automation API and ChatGPT, Codeo’s app allows you to use plain language to issue commands in Inventor. You can type something as simple as “Close Inventor,” or voice complex modeling commands such as “Create a cylinder with a diameter of five centimeters and a height of ten centimeters.” And voilà, the model is born. You can then move and adjust the model, for example, telling it to “Zoom in on the top of the cylinder.”  

The ChatGPT engine processes natural language, making it easy to ask for what you need and bypass complex processes for filtering and manipulating models. Collaboration improves, too. Tell the app to “Add a note to the cylinder to ‘Check for strength,’” and ChatGPT would create that annotation in the model. 

Want to learn more? Tell Siri to pull up Codeo’s website

Related Learning: 
It’s Not Too Late to Automate: Using Forge Design Automation for Inventor  
The team from Codeo shows you how to automate design tasks at scale in the cloud using the Design Automation API for Inventor, freeing your team members from drudgery so they can spend more time being creative. 

Close quality control gaps on megaprojects with AI 

It’s not humanly possible to verify every part of a construction project against the plan. Buildings are too complex. Today, teams use spot checks, surveys, and laser scanning to identify issues. These manual processes mean 90% to 95% of a project goes unchecked for errors and gaps, according to Naska.AI’s estimates.  

But what’s impossible for humans is possible for computers. Naska.AI builds AI tools and algorithms to exponentially increase your ability to manage quality and timelines and prevent errors. 

Naska.AI uses Autodesk Platform Services APIs to extract, process, and visualize model data in the cloud, where advanced technology can be applied to: 

  • Compare laser scan data to models 
  • Automate identification of critical issues early, reducing rework 
  • Highlight deviations from models to improve coordination between trades 

These processes happen through an intuitive workflow that’s built on top of Autodesk software, so you don’t have to learn to navigate a new interface. 

In one use case, a builder had taken over a partially built megaproject after the previous contractor went bankrupt. The customer wanted to quickly compare BIM and shop drawings to reality. Using Naska.AI, they analyzed more than 100,000 structural elements in a month, flagged 94 critical issues, and mitigated $1.5 million in risk. 

Learn more about Naska.AI.   

Related Learning: 
Reality capture process from cloud platform using Forge to merge BIM model 
Learn a workflow that takes point cloud data from LIDAR into the cloud with APS, then merge it with model data in Revit for comparison, creation of a digital twin, and more. 

Wide-ranging AI-enhanced workflows for building and product design 

CCTech has been creating 3D configurators and automation workflows for more than 15 years. As generative design has matured, so too have CCTech’s ambitions to bring AI to 3D workflows for building and product design.  

The key was taking CAD and BIM to the cloud. CCTech is using APS as the foundation for a range of AI-enhanced applications. The APS Viewer allows designs to respond to AI predictions in real time, while Design Automation APIs provide automation workflows between Inventor, Revit, AutoCAD, and the cloud. 

The resulting solutions include: 

  • Occupancy detection using CCTV: Building Management Systems (BMS) automate HVAC systems to keep people comfortable and optimize energy use, but the systems have limited data inputs. Sensors can tell you how many people are present, but they can’t tell you exactly where they’re located, for example. To bring more precision to energy management, CCTech built AI-based computer vision using CCTV footage that allows the system to detect occupant positioning and activity. This solution integrates with the building’s digital twin, powered by APS.  
  • AI-powered valve design: First, CCTech built a configurator for valve designs. Then its programmers layered on AI to create an even faster shortcut for designers. The result is an app that allows the designer to set the valve performance parameters and in return receive a set of five potential designs that meet those goals. With a simple interface of sliders for each parameter, even non-professionals can use the tool and receive their designs within minutes. 
  • Autonomous mesh segmentation: Designers often lose their original design files, leaving them with lower-grade STL or OBJs. CCTech uses AI to restore the integrity of the source file. Upload a file, and the AI model classifies each triangle into a surface type like planar, spherical, torus, etc. As CCTech builds out the technology further, it will allow you to recreate a feature-rich CAD model. 
  • 3D object detection: CCTech is using machine learning to turn low-data CAD models into data-rich BIM models. Machine learning is already used to classify single objects. But CCTech has expanded this capability to detect multiple objects. You can enhance the CAD model with detailed data that may have been missing, such as window dimensions, equipment, or materials.  

Learn more about the solutions CCTech builds and about their own SaaS CFD platform

Related Learning: 
Why and how to build FDX Connectors?  
Data Exchange helps break the barriers of data siloes developed within an organization. The team from CCTech demonstrates how the Data Exchange API liberates data from proprietary formats. allowing seamless integration with desktop apps, web apps, mobiles apps, and more. 
Bridging Revit & Inventor using Autodesk Data Exchange to make configurator
Data Exchange gives you the control to share the right information with the right person at the right time, regardless of industry or application. The team from CCTech shows you how to use the Data Exchange service to connect different CAD software. 

Learn more about the possibilities of AI and the cloud on the Autodesk Platform Services page of the AU website.