Description
Key Learnings
- Learn about using Smart Blocks to improve drafting efficiency.
- Learn how to integrate Smart Blocks features into your workflows to reduce the time and effort needed to cleanup drawings.
- Learn how to use Markup, Import, and Assist to drastically reduce the time it takes to incorporate feedback into an AutoCAD drawing.
Speakers
- AMAnum MithaniAnum's career began in the engineering and construction industries. She is a graduate of the University of Waterloo, holding a Bachelor's in Civil Engineering. Her professional experience includes four years as a building science consultant, where she focused on projects involving low-energy new construction and restoration of heritage buildings. Anum completed an MBA and a Masters in Engineering in Design Innovation from Northwestern University. During this time, she interned with the AutoCAD Desktop Product Management team at Autodesk, and upon graduating she returned to the team full-time. In her current role as a Product Manager for AutoCAD for Desktop, Anum contributes to the customer experience of the Smart Blocks Search and Convert and Object Detection workflows in AutoCAD. She is also working on strategies and features for insights in AutoCAD.
- ARAnand RajagopalAnand is the Head of AI Development for AutoCAD at Autodesk, where he has spent the last decade exploring the intersection of AI and Architecture, Engineering, and Construction (AEC). With a passion for harnessing the potential of data and AI, Anand has dedicated his career to improving the connectivity of data across the entire construction lifecycle. He has developed innovative products that cater to both superintendents in the field and drafters in the office. Anand holds two patents for his research focused on enhancing job site safety through data utilization.
ANUM MITHANI: Good afternoon and welcome to today's product demo. We really appreciate all of you choosing to tune in to today's presentation. We've got a pretty exciting topic to cover here. We're going to explore Autodesk AI features within AutoCAD, namely Smart Blocks and Markup Import and Assist.
But before we get into the demo, we'd like to introduce ourselves. I'm Anum Mithani, Product Manager on the AutoCAD Desktop Team. I work on the desktop workflow experience for some of our ML features, as well as our AutoCAD insights strategy. And with me today I have.
ANAND RAJAGOPAL: Hi. My name is Anand. I'm an Engineering Manager within AutoCAD. I lead the AutoCAD AI and ML development teams. Happy to be here.
So we're really coming to you today with just one goal. Our goal is to really introduce you to what the Autodesk AI features within AutoCAD look like and how can you leverage them to better increase your own productivity. Let's dive, right [AUDIO OUT]
We've all seen the AI kind of buzz around us. We felt it. It's been there now almost two years. And it's gone beyond just ChatGPT. That's still what's spoken about the most, but it's making its circles within the AEC realm. You see it every day talking about the productivity enhancements in our industries.
And we want to really tell you what we're doing within AutoCAD. But before we get there, we want to just maybe take a minute to level set. What is AI? Next slide, please.
So despite the current attention, I isn't really a new term. It was coined in the 1950s. It's really a broad area of computer science research.
It focuses on tasks that need humans and computers to interact on tasks which require usually human-level intelligence, but it's not restricted to that. Within this broad umbrella of AI, machine learning, or ML is a subset of techniques. It's just a set of algorithms that rely on today's advanced processing capabilities to sift through large data volumes and discover patterns.
It's really an implementation-- one kind of implementation for what AI is today. And it's kind of really synonymously used with AI. AI, ML you kind of see that and you can kind of really interpret them the same way. I want to take a second to call on a specific subset of even machine learning algorithms, the generative AI algorithms.
I mean, these are the ones which are really hard. People are talking about it everywhere. These are basically a specific set of algorithms which can create new content. ChatGPT for text, Midjourney and the likes for images. This isn't something we are talking about today, but if you have more questions, you can come meet us at the booths.
It's something we can definitely talk to you more about. But this is also honestly as technical as I want to get today, because our goal isn't to teach you more about what is AI, but really what can we give you within the products leveraging AI. So can we go on to the next slide, please?
What AI can do for you. So in Autodesk, we look at AI in three different ways. Build. First one is automation. This is to reduce the repetitive, tedious work and increase your own time for the kind of work that you matters most to you.
The next one is augmenting. Can we augment your everyday workflows with AI to make them better partner with you to keep you in control, but again, free up your time to do the things that matter most? And finally, analyze.
This is to navigate the large mountains of information that comes today to help you make better decisions. And what you're going to do in the subsequent presentation is talk about how all of these workflows manifest themselves within AutoCAD today. With that, I'm going to pass it on to Anum to walk us through some examples.
ANUM MITHANI: Thank you. So now that we've got a solid foundation on what is AI, let's see how Autodesk AI can help customer workflows. We'll be using an example to highlight the key Autodesk AI features within AutoCAD.
And here is our lovely lighting designer, Riley. She's teaming up with an architect on an exciting new project. Today, we'll be following along on her journey and get a sneak peek into her workflow as she reviews a background drawing, designs her lighting plan, adds and replaces lighting elements, and gets feedback from her manager. And guess what? She's doing it all with a little bit help from Autodesk AI. So let's dive right in.
Riley just got a new drawing from the architect. It's a detailed floor plan of a two-story house. The drawing shows walls, doors, furniture, and even some light fixtures that the architect has already specified.
Since Riley imported this drawing from a different program, all of the objects within the drawing are exploded, meaning that there are no blocks in the drawing. It's a little bit of a puzzle now, but Riley's ready to figure it out. It's time for some much needed drawing cleanup.
Riley's reviewing the architect's drawing and notices that some light fixtures are already included. She spots a wall light and sees another one right next to it. Riley decides to convert all of the wall lights into blocks for easier management. Normally, she would have to painstakingly select the geometry, create the block, and then manually search through the entire drawing to find all similar geometries, deleting and replacing them one by one.
It's a tedious process, and often, objects will get missed. But then, Riley remembers the Smart Block Search and Convert feature within AutoCAD. This handy tool can automatically find all matching geometries. She selects the wall light geometry for conversion and types in the BCONVERT command, and sees all of the exact matches that are in the drawing.
When reviewing the found instances, Riley can easily remove or add back any of the found instances into her selection. Ready to proceed, she presses enter to convert, and the Convert to Block Palette opens up. This palette displays a total number of instances found on the left. As well, there is a checkbox filter that will automatically remove any found instances that are already blocks.
But because Riley already knows everything in the drawing is exploded, she doesn't need to use this filter. Riley can choose to convert to an existing block from her library, or to create a new one. Since this wall light is unique to the architect, she decides to create a new block and names it arc wallet.
AutoCAD will automatically select the geometry center as its base point. But Riley wants a base point to be where the fixture meets the wall. She clicks on Select Base Point and picks the base point directly within the Canvas. Finally, she can also define the layer that the new block instances should be inserted onto. She selects her predefined lighting layer, ensuring that all the blocks are on the correct layer.
With everything set, she completes the conversion with ease, transforming the scattered wall lights into organized blocks, making her drawing much more manageable. We can see that all of the found instances in the drawing are now the new block.
Feeling accomplished with those wall lights, Riley notices that the architect has also included some pendant lights in the drawing. Once again, to save time from scouring through the drawing manually, she uses the Search and Convert feature. She selects one pendant light, types in BCONVERT and AutoCAD quickly finds and highlights all of the instances of the same geometry. With the instances selected, she proceeds to convert them.
Riley knows that her firm commonly uses pendant lights and has a block stored in their block library. So she opts to convert the instances into an existing block. AutoCAD suggests similar blocks for her to choose from, thanks to machine learning algorithms that use geometry and naming to find similar blocks that already exist within your library.
Once again, she can choose the right layer for the block. She clicks on the lighting layer to keep everything organized and goes ahead to convert. Riley's now saved a significant amount of time and effort by using the Search and Convert feature. The ability to quickly convert repetitive geometry into blocks minimizes redundancy in the design and provides wider options for organizing drawings and having more performant drawings.
With the wall lights and pendant lights efficiently converted into blocks, Riley's drawing is now much more streamlined and manageable. She can focus on the finer details of the design, knowing that the tedious part of the process has been handled with ease.
ANAND RAJAGOPAL: Thank you, Anum, for giving us an overview. I'm just going to do a quick recap of what we just saw there. With Search and Convert, we saw Riley efficiently identify and convert unblocked lighting elements. It helped us search the drawing to select all matching geometry, convert the geometry into an existing block, or new blocks, in just one easy step.
BCONVERT, that's the command, let's Riley scale her work but still keeps her in control. It allowed her to define custom base points, select appropriate layers, ensuring that the converted geometries fit seamlessly into her design. We're going to follow along with Riley and see what's her next move. Back to you, Anum.
ANUM MITHANI: Awesome. So with the wall lights and pendant lights efficiently converted into blocks, Riley decides to continue with her drawing cleanup and tackle the rest of the drawing. Remember, all of the objects came in exploded. She wants to block everything else in the drawing quickly to realize blocks performance gains.
Normally this process would take hours, but Riley would have to go through and identify each object like a piece of furniture, manually selecting each piece of geometry that makes up that object. This can be super time consuming and tedious if the object is small, detailed, or not easily selectable using a selection window. Then she has to convert it into a block and manually find all of the similar objects to delete and replace them with the new block.
Repeating this over, and over, and over for each type of object would be exhausting. But Riley is savvy and knows that Smart Blocks Detection uses machine learning to scan her drawing and identify objects that can be converted into blocks automatically.
She heads to the ribbon and clicks on the Detect Button to launch the detection process. The palette opens and she can see it's working to scan the drawing for potential blocks. The best part? It's working in the background, so Riley can close the palette to regain full screen access and continue working on other important tasks.
A notification soon lets her know that the detection results are ready. From the notification, Riley sees that 17 sets of objects have been found, covering a total of 48 instances. She clicks to review, which takes her to Detection Review mode. At the top, the device-- the visor displays the 17 sets that were found and she can use arrows to flip between the results.
Each set represents a group of similar detected objects that can be converted into instances of a single block. The visor has other helpful tools. The Convert button. When Riley is ready to convert the set into a block, she can simply click on this button.
Next is the Report an Error button. If there's an issue with the detected set, Riley can report an error. By reporting errors, she contributes to improving the overall performance and accuracy of the machine learning service.
The Primary Instance button. This button allows Riley to select a new primary instance. Within each set, the primary instance, outlined by a blue border, serves as a block definition when converting a new block.
Alternatively, if she chooses to convert into an existing block, the primary instance is used to define the insertion scale and rotation for the block. And finally, Remove Instances. Riley can use this tool to remove any instances within the detected set that should not be converted with the rest.
Riley then spots a detected set that contains store objects. She wants to convert all of the interior doors, but not the exterior ones into a block. By using the handy Remove Instances tool, she can carefully remove the exterior doors from the selection set, ensuring that only interior doors are included.
When she clicks to Convert, the Convert to Block palette will open up, showcasing all the helpful capabilities she saw with Search and Convert. She decides to convert the door into an existing block in her library and chooses a half-closed door from the machine learning based suggested blocks.
Riley selects the appropriate lighting layer, ensuring everything is organized correctly. When she presses Convert, AutoCAD gives her the option to accept the placement of the existing block, or to adjust it as needed. She moves the block into the correct position. Then, she scales it to fit and accepts the changes.
Now, we can see that all of the interior doors have been changed into the new block, thanks to Riley's smart use of the tools at her disposal. The drawing has become more organized and performant by the minute. Riley continues through the detection results, swiftly creating new blocks en masse. Each set she reviews brings her closer to a more organized and performant drawing.
However, the next set she encounters has an error. No problem. Riley clicks on the Report Error button. She knows that while detection is best at identifying architectural objects on floor plans, it's not always perfect in this release. By reporting errors, she's helping to continuously improve detection results over time.
She continues on, empowered, knowing that her drawing is as accurate as possible and leveraging tech to complement her expertise. As Riley continues, she feels a sense of accomplishment. The once chaotic and exploded objects have been transformed into well-defined blocks, making the entire layout so much more manageable.
She takes a moment to appreciate how much time and effort she saved. Instead of manually searching for and converting each object, she's harnessing the power of Smart Block Detection to do all the heavy lifting. This has allowed her to quickly move on and focus on her next important task of designing the lighting plan, where her expertise will truly shine.
ANAND RAJAGOPAL: Thank you, Anum. Way to go, Riley. We earlier just saw her use the BCONVERT command, and that was to find matching lighting elements. This time, we saw her use the Detect command to automatically scan and identify all unblocked objects that can then be converted into blocks in her exploded drawing.
Under the hood, Detection uses machine learning to determine if a particular group of geometric elements should be blocked, or are usually blocked. This, again depends on the context of the drawing. So it's something which is always going to learn from how people engage with the particular command and how they use it.
This was a great way to clean up the drawing. The AI today, as Anum mentioned, is better at certain objects and is always getting better. And it should continuously learn and improve with subsequent releases. With that, I'm going to pass it back to Anum to move on and show us what the lighting plan looks like.
ANUM MITHANI: Awesome. So as Riley progresses through her lighting design, she has to place blocks from her blocks library into the drawing. Normally, this can take some time. She has to make sure that the scale and rotation for each block makes sense for the floor plan.
But as she continues on with her work, the Smart Blocks Placement Engine begins to learn from how the existing block instances have been placed in the drawing. This intelligent engine starts to infer the optimal placement and scale for the next instance of the same block. When Riley goes to insert another wall light, the engine provides placement suggestions near similar geometries where that block has been placed before. Impressively, it autorotates the wall light to align with the wall.
This feature is a gamechanger for Riley. It significantly speeds up the process of placing blocks, ensuring that they're correctly oriented and positioned without the need for constant manual adjustments. With each block placement, the engine becomes more accurate, learning from Riley's design choices and making smarter suggestions. This not only saves time, but also reduces the chance of errors ensuring a more consistent and professional layout.
ANAND RAJAGOPAL: Thank you, Anum. We just saw the Smart Blocks placement engine learn on the go as Riley continued her work within AutoCAD. It inferred the optimal placement, the scale, the rotation for the new instances. And this was based on how the previous blocks have been placed, or could have been-- it provides some smart suggestions Riley, based on what she's looking to do. It could automatically recommend a rotation or multiple placement of something in the right spot and the right place.
And like Anum said, this helped to go much faster on the overall lighting plan. It helped to ensure a more consistent layout, consistent design, and help reduce the chance of errors. Riley still makes all the decisions, but she's just got a smart tool in her pocket to help her go faster. With that, let's move on to our last Smart Blocks workflow.
ANUM MITHANI: Awesome. So Riley's now got a lot of her lighting plan done, but she needs to fine tune. She identifies some hallway lights that need to be replaced with a different type of light symbol. Instead of deleting the blocks she's already placed and inserting the correct light symbol, she knows she can leverage Smart Block's replacement.
She selects all the instances of the hallway lights that she wants to replace and types in the BREPLACE command. AutoCAD suggests similar blocks for her to choose from. She can also replace the selected blocks with another block she specifies, either from a drawing or from a list of recently inserted or created blocks.
The beauty of this feature is that the replaced block retains the scale, rotation, and attribute values of the original block. She selects the correct block from the suggested block list, and instantly all of the hallway lights are replaced with the new symbol. This seamless replacement process saves Riley a significant amount of time and effort, ensuring that consistency and precision that she needs in her drawing.
Her drawing's now much more polished and refined, thanks to the powerful features of Smart Block Search and Convert, Detection, Placement, and Replacement. The ability to quickly replace blocks without losing their properties has made her job so much easier.
ANAND RAJAGOPAL: Thank you. So we saw the Smart Block Replacement workflow and the BREPLACE workflow-- or the BREPLACE command, sorry. This used to determine good replacement candidates for the given blocks, and this was based on the blocks present in Riley's local library. So it works great with what you already have. It just needs you. It just understands what is a good replacement, either based on prior use, based on how it's defined, or just what it represents.
It still retains the same scale, the rotation attribute values of the original blocks. This ensured that Riley could update the plan seamlessly, without the need of any further adjustments. Yeah. And I think with that, we've seen a whole host of Smart Blocks workflows, but we've got actually one more AI-based workflow that we want to show you. Back to Anum.
ANUM MITHANI: OK. So now we know the lighting design is complete, Riley emails a PDF of the drawing to her manager for review. Her manager is pretty old school, so she's printed out her PDF and made handwritten markups, leaving the paper copy on Riley's desk.
Typically, Riley would have to keep moving her eyes from her desk to her screen, trying to find all of the places where edits need to be made and keeping track of which ones she's already incorporated as she works through the list. But Riley knows about Markup Import and Assist, which can make this process much smoother. She takes a photo of the markup, drawing on her phone and saves it to her computer.
Then she heads to the Collaborate tab in the ribbon and clicks on Markup Import. Riley selects the photo from her files to import the markup image. Markup Import overlays a markup file on top of the drawing in the Trace workspace. Imported markups are automatically aligned with the existing drawing. Sometimes, the markups don't align perfectly and require manual adjustments. Riley can move, align, rotate, or scale the markup if needed.
Once the markup is properly aligned, Riley switches to the Drawing workspace. Here, Markup Assist comes into play by identifying text in the imported markup and highlighting it with blue boxes. She begins by editing the lighting notes. By clicking on the blue box around her manager's handwritten addition, Markup Assist converts a handwritten message in a text. Riley can then easily append this addition into her notes.
Markup Assist can also identify strikethroughs. Riley uses this feature to erase existing text and replace it with the corrected markup, ensuring that her drawing reflects the latest updates from her manager. She can also insert new text directly from the markup by selecting it and inserting it as an MText. This is incredibly useful for adding new notes and annotations.
Next, we'll see that Markup Assist can also interpret instructional texts. When Riley clicks on the blue box around the Remove Markup, it triggers the Erase command, allowing her to remove specified objects. Once she completes the task, she can fade the markup, indicating that the instruction has been addressed.
Similarly, when Riley clicks on the Move Markup-- oh, wait. She's still fading markups. Awesome. Similarly, when Riley clicks on the Move Markup, the Move command is triggered, enabling her to select and relocate objects as directed. This automation saves her a significant amount of time and ensures precision in following her manager's instructions. Again, we see here fading the markup once it's been addressed.
Riley can also insert text as MLeader with Markup Assist. She clicks on the markup, and, if needed, she edits the text. Sometimes, her manager's notes can be a little bit messy. She then selects Insert as MLeader and places it precisely where it's required. She then fades the arrow markup and displays the Mleader arrow behind.
With these tools at her disposal, Riley quickly incorporates all of her managers mark ups. She's efficiently handled the addition of new text, removal of objects, and relocation of elements, all with the help of Markup Import and Assist. Now she's ready to send the updated drawing back to the architect.
She reflects on how much time she saved using Markup Import and Assist, along with all of the Smart Blocks features. These tools have not only made her workflow more efficient, but again, has ensured that her drawing is accurate and professional. She feels confident and prepared that if any future revisions or updates come in, she has the right tools to handle them with ease. With these advanced features, she can continue to work smarter and not harder to achieve her design goals more efficiently.
ANAND RAJAGOPAL: Thank you, Anum. We saw Riley use Markup Import and Assist, two commands to streamline the process of incorporating feedback. Riley used Import to conveniently overlay the photo and the information in that photo on top of a drawing in the Trace workspace.
Once this data is in Trace, this is when Markup Assist kicks off. It helps-- it smartly helped her make the changes needed. It identified and highlighted the handwritten text on the photograph, used OCR, or optical character recognition, to quickly convert that under the hood into text, which the computer can then work with, or edit.
The comments are now ready in the clipboard. They can be quickly incorporated into either MLead or MText. Everything is just with one click.
It can go further. It can understand the intent of some of this feedback and helpfully trigger the right workflows. We saw examples of Remove. We saw examples of Move, or Erase. It can even identify which objects need to move. This whole process augments her current collaboration process and makes every iteration go a little bit faster.
Thank you. With this, we finished showing you four different workflows. Now Anum is going to walk you through how you can actually access these features today and where they're available. Back to you, Anum.
ANUM MITHANI: OK. So now that you've seen Riley's workflow, we hope that you guys are ready to enhance your own workflows. Let's dive right in on how to access these powerful features, explore detailed tutorials, and share your feedback to help us improve.
All of the awesome Autodesk AI features within AutoCAD that we've talked about today are ready for you to use in our latest release, AutoCAD 2025. You can also find a subset of these features in earlier releases, namely AutoCAD 2023 and 2024. Some of these features, Placement and Search and Convert, are also available in AutoCAD LT.
To get the most of these amazing features, here are some helpful tips. First, you can learn more. You can always visit help.autodesk.com to dive deeper into these features and their applications. You can also access help articles by clicking on the Learn More link within the product to find detailed information about specific features as you encounter them.
And we also have Autodesk Assistant right within the product to help answer any of your questions. This chat bot will provide you AI-recommended solutions for your inquiries and connect you with an agent, if it's needed.
Provide feedback. Help us make detection even better. Click on the Provide Feedback in the Detection palette. This will direct you to a survey where you can share your valuable input.
And finally, report errors. If you encounter any issues, you can report errors in the Detection Review visor. This tool displays a list of potential errors within the current set of detected objects. Reporting these errors contributes to our ongoing improvement of the machine learning service's performance and accuracy.
It's a collaborative effort to refine the AI, ensuring it becomes even more accurate and helpful in the future. While you may not see that the results are perfect all the time, you have the control and the tools that you need to make it work best for you. By leveraging these resources, you can enhance your experience within AutoCAD and help us improve our AI features. There's also additional classes at AU that you can visit to learn more about these features.
Thank you for joining us today and exploring how Autodesk AI features can transform your workflow, saving you time and enhancing your design process. We'd also like to share an opportunity to connect with the AutoCAD team and have a say on our next releases. Scan the QR code to learn more about our inside the factory event. Thank you.
ANAND RAJAGOPAL: Thank you.