[{"data":1,"prerenderedAt":954},["ShallowReactive",2],{"{\"resolve_relations\":\"reference.source\",\"version\":\"published\"}resources/blog/why-ai-in-manufacturing-needs-to-understand-cad-data":3},{"data":4,"headers":931},{"story":5,"cv":914,"rels":915,"links":930},{"name":6,"created_at":7,"published_at":8,"updated_at":9,"id":10,"uuid":11,"content":12,"slug":905,"full_slug":906,"sort_by_date":907,"position":908,"tag_list":909,"is_startpage":294,"parent_id":910,"meta_data":97,"group_id":911,"first_published_at":912,"release_id":97,"lang":794,"path":97,"alternates":913,"default_full_slug":97,"translated_slugs":97},"Why AI in Manufacturing Needs to Understand CAD Data  ","2026-06-30T14:16:58.075Z","2026-06-30T18:14:28.138Z","2026-06-30T18:14:28.169Z",193083056647257,"4029a7fb-98cc-407f-9552-b6032b41a07c",{"_uid":13,"body":14,"date":745,"fold":746,"unit":818,"intro":819,"title":807,"author":826,"sidebar":827,"category":886,"end_date":17,"location":17,"metadata":893,"component":898,"thumbnail":899,"past_event":294,"hide_author":294,"english_only":294,"hide_sidebar":294,"redirect_url":17,"hide_from_feed":294,"show_date_range":294,"is_external_page":294,"open_graph_image":903,"external_page_url":17,"include_open_graph":131,"block_search_engines":294},"015c4825-95f3-4f19-abd8-1dcd8091b8a2",[15,19,132,194,279,295,373,463,568,582,593,604,615,626,635,644,646,654,660,671,682,701,717],{"_uid":16,"height":17,"component":18},"fc3725b4-7fbc-4e55-9c67-90291430fe64","","element-spacer",{"_uid":20,"body":21,"alignment":17,"component":130,"is_full_width":131},"d6494edd-4bb7-4c9f-ba48-c2d26b77a7c2",{"type":22,"content":23},"doc",[24,48,56,74,82,109],{"type":25,"attrs":26,"content":28},"paragraph",{"textAlign":27},"left",[29,36,43],{"text":30,"type":31,"marks":32},"After leaving its mark on seemingly everything else, artificial intelligence is finally touching engineering workflows, improving how teams design, simulate, and optimize products. For manufacturing applications developers, the next steps for AI innovations requires overcoming a massive obstacle: ","text",[33],{"type":34,"attrs":35},"textStyle",{"color":17},{"text":37,"type":31,"marks":38},"How do you integrate CAD data into AI workflows?",[39,41],{"type":34,"attrs":40},{"color":17},{"type":42},"bold",{"text":44,"type":31,"marks":45}," ",[46],{"type":34,"attrs":47},{"color":17},{"type":25,"attrs":49,"content":50},{"textAlign":27},[51],{"text":52,"type":31,"marks":53},"CAD data sits at the center of modern manufacturing workflows. It carries geometry, structure, product knowledge, and engineering intent from design throughout analysis, production, inspection, and maintenance. It is also highly proprietary and far more complex than the text and image data most AI tools are built around. ",[54],{"type":34,"attrs":55},{"color":17},{"type":25,"attrs":57,"content":58},{"textAlign":27},[59,64,70],{"text":60,"type":31,"marks":61},"To create the next generation of manufacturing applications, developers need to connect machine learning workflows to real CAD data without compromising IP, losing the engineering context that makes the data valuable, or creating outputs that aren’t useful downstream. In summary, ",[62],{"type":34,"attrs":63},{"color":17},{"text":65,"type":31,"marks":66},"manufacturing applications need AI functionality built around and explicitly for CAD data.",[67,69],{"type":34,"attrs":68},{"color":17},{"type":42},{"text":44,"type":31,"marks":71},[72],{"type":34,"attrs":73},{"color":17},{"type":25,"attrs":75,"content":76},{"textAlign":27},[77],{"text":78,"type":31,"marks":79},"In this post, we outline what makes CAD data and AI such a problem for developers and what needs to happen to practically address these issues.  ",[80],{"type":34,"attrs":81},{"color":17},{"type":25,"attrs":83,"content":84},{"textAlign":27},[85,90,104],{"text":86,"type":31,"marks":87},"Finally, we share ",[88],{"type":34,"attrs":89},{"color":17},{"text":91,"type":31,"marks":92},"HOOPS AI",[93,100,103],{"type":94,"attrs":95},"link",{"href":96,"uuid":97,"anchor":97,"target":98,"linktype":99},"https://www.techsoft3d.com/developers/products/hoops-ai/",null,"_blank","url",{"type":34,"attrs":101},{"color":102},"#467886",{"type":42},{"text":105,"type":31,"marks":106},", the CAD-based framework designed specifically for connecting engineering data with machine learning, can help manufacturing application developers overcome these issues.  ",[107],{"type":34,"attrs":108},{"color":17},{"type":110,"content":111},"blockquote",[112],{"type":25,"attrs":113,"content":114},{"textAlign":27},[115,117,121,123,129],{"text":116,"type":31},"Check out our blog post covering our conversation with Digital Engineering 24/7 Senior Editor ",{"text":118,"type":31,"marks":119},"Kenneth Wong",[120],{"type":42},{"text":122,"type":31}," to learn more about ",{"text":124,"type":31,"marks":125},"how AI is already transforming engineering, design, and simulation workflows.",[126],{"type":94,"attrs":127},{"href":128,"uuid":97,"anchor":97,"target":98,"linktype":99},"https://www.techsoft3d.com/resources/blog/ai-impacting-engineering-design-simulation/",{"text":44,"type":31},"copy",true,{"_uid":133,"body":134,"alignment":17,"component":130,"is_full_width":131},"f2d2f96e-8fff-4e49-b1c5-d6b61a168d50",{"type":22,"content":135},[136,143,151,159,167,186],{"type":137,"attrs":138,"content":140},"heading",{"level":139,"textAlign":27},2,[141],{"text":142,"type":31},"Challenge 1: The Most Useful CAD Data Usually Cannot Leave the Company ",{"type":25,"attrs":144,"content":145},{"textAlign":27},[146],{"text":147,"type":31,"marks":148},"Most manufacturing CAD data is highly confidential. It may contain product geometry, design intent, customer requirements, production knowledge, supplier constraints, and other details that should not be exposed outside the organization. ",[149],{"type":34,"attrs":150},{"color":17},{"type":25,"attrs":152,"content":153},{"textAlign":27},[154],{"text":155,"type":31,"marks":156},"This is the first major issue for integration with AI tools. AI systems are most useful when they can learn from or operate on data that reflects the real workflows they are meant to support, and specialized data you would want to use to train a model is the info manufacturers are least willing or able to share. The unique nature of manufacturing data also means publicly available data is likely not sufficient. ",[157],{"type":34,"attrs":158},{"color":17},{"type":25,"attrs":160,"content":161},{"textAlign":27},[162],{"text":163,"type":31,"marks":164},"General-purpose AI tools are powerful in part because they can learn from enormous public datasets. That same foundation does not exist for proprietary CAD data.  ",[165],{"type":34,"attrs":166},{"color":17},{"type":25,"attrs":168,"content":169},{"textAlign":27},[170,175,181],{"text":171,"type":31,"marks":172},"For developers, this creates a clear requirement for their applications: ",[173],{"type":34,"attrs":174},{"color":17},{"text":176,"type":31,"marks":177},"manufacturing AI needs to work within the systems and environment where the CAD data already lives",[178,180],{"type":34,"attrs":179},{"color":17},{"type":42},{"text":182,"type":31,"marks":183},". It needs to support machine learning workflows inside controlled environments, using the company’s own engineering data without unnecessarily exposing sensitive IP. ",[184],{"type":34,"attrs":185},{"color":17},{"type":25,"attrs":187,"content":188},{"textAlign":27},[189],{"text":190,"type":31,"marks":191},"This is one of the core ideas behind HOOPS AI. HOOPS AI gives developers a framework for applying machine learning to CAD data in controlled engineering environments. Instead of treating CAD as generic files to be sent somewhere else, it supports workflows built around the realities of proprietary engineering data. ",[192],{"type":34,"attrs":193},{"color":17},{"_uid":195,"body":196,"alignment":17,"component":130,"is_full_width":131},"bba5f3de-272c-4047-9313-0fac157cffd0",{"type":22,"content":197},[198,203,221,229,237,255,263,271],{"type":137,"attrs":199,"content":200},{"level":139,"textAlign":27},[201],{"text":202,"type":31},"Challenge 2: CAD Data Cannot Be Read Like Text or Images ",{"type":25,"attrs":204,"content":205},{"textAlign":27},[206,211,217],{"text":207,"type":31,"marks":208},"Even with security concerns overcome, manufacturing application developers run into the next issue: ",[209],{"type":34,"attrs":210},{"color":17},{"text":212,"type":31,"marks":213},"CAD data is not naturally readable by AI.",[214,216],{"type":34,"attrs":215},{"color":17},{"type":42},{"text":44,"type":31,"marks":218},[219],{"type":34,"attrs":220},{"color":17},{"type":25,"attrs":222,"content":223},{"textAlign":27},[224],{"text":225,"type":31,"marks":226},"Obviously, a CAD file is not simply an image of a part. The file contains B-Rep geometry, topology, dimensions, numerical values, structured text, assembly information, feature trees, metadata, and other engineering information. While that structure and data richness are what make CAD useful, it is also makes CAD difficult for general-purpose AI systems to interpret directly. ",[227],{"type":34,"attrs":228},{"color":17},{"type":25,"attrs":230,"content":231},{"textAlign":27},[232],{"text":233,"type":31,"marks":234},"Engineers interact with a CAD model only after their chosen software reads the file, interprets the geometry, renders the shape, and provides tools to inspect it. Machine learning systems need their own version of that translation layer. The geometry and structure of the model need to be converted into representations usable for model training, inference, and evaluation. ",[235],{"type":34,"attrs":236},{"color":17},{"type":25,"attrs":238,"content":239},{"textAlign":27},[240,245,251],{"text":241,"type":31,"marks":242},"For developers, this ",[243],{"type":34,"attrs":244},{"color":17},{"text":246,"type":31,"marks":247},"means practical AI infrastructure for engineering applications needs CAD access, CAD interpretation, conversion into machine-learning-ready formats, and a way for humans to review results against the original geometry.",[248,250],{"type":34,"attrs":249},{"color":17},{"type":42},{"text":44,"type":31,"marks":252},[253],{"type":34,"attrs":254},{"color":17},{"type":25,"attrs":256,"content":257},{"textAlign":27},[258],{"text":259,"type":31,"marks":260},"HOOPS AI is built to support that workflow. The toolkit can read CAD data and convert it into tensor-ready representations for machine learning. It also includes visualization capabilities so users can inspect CAD data and evaluate machine learning results in context. ",[261],{"type":34,"attrs":262},{"color":17},{"type":25,"attrs":264,"content":265},{"textAlign":27},[266],{"text":267,"type":31,"marks":268},"That review step matters. AI will make mistakes, and engineering teams need a way to verify outputs before they influence design, manufacturing, or inspection decisions. ",[269],{"type":34,"attrs":270},{"color":17},{"type":25,"attrs":272,"content":273},{"textAlign":27},[274],{"text":275,"type":31,"marks":276},"HOOPS AI also supports CAD-specific machine learning approaches such as UV-Net, giving developers a practical starting point for working with BRep geometry and topology. ",[277],{"type":34,"attrs":278},{"color":17},{"_uid":280,"title":281,"contents":282,"component":283,"description":284,"exclude_schema":294},"1e0aa11e-1541-4265-b59a-28dc36f3adef","Technical Note: What is UV-Net? ",[],"element-faq",{"type":22,"content":285},[286],{"type":25,"attrs":287,"content":288},{"textAlign":97},[289],{"text":290,"type":31,"marks":291},"UV-Net is a machine learning architecture designed for BRep CAD data. It combines CNN-based learning from surface geometry with GNN-based learning from topology, helping models understand both the shape of a part and how its faces, edges, and other structures relate to each other. ",[292],{"type":293},"italic",false,{"_uid":296,"body":297,"alignment":17,"component":130,"is_full_width":131},"d08cdd92-6d08-4583-9b12-9d4516613573",{"type":22,"content":298},[299,304,309,314,319,324,329,334,339],{"type":25,"attrs":300,"content":301},{"textAlign":97},[302],{"text":303,"type":31},"HOOPS AI supports UV-Net, giving developers a practical starting point for experimenting with BRep-based machine learning without building that foundation from scratch. ",{"type":137,"attrs":305,"content":306},{"level":139,"textAlign":97},[307],{"text":308,"type":31},"Challenge 3: AI Results Need to Fit Real Engineering Workflows ",{"type":25,"attrs":310,"content":311},{"textAlign":27},[312],{"text":313,"type":31},"The third challenge for developers comes after AI tool they’ve created produces a result. ",{"type":25,"attrs":315,"content":316},{"textAlign":27},[317],{"text":318,"type":31},"In manufacturing, a useful output needs to fit into the next step of a complex, customer-informed workflow. That may mean design review, simulation, CAM, inspection, documentation, data exchange, or further editing by an engineer. ",{"type":25,"attrs":320,"content":321},{"textAlign":27},[322],{"text":323,"type":31},"This is where many AI demos can become misleading. A visually plausible 3D result may be impressive, but manufacturing teams need more than a shape on screen. They need results that can be checked, trusted, and used in the systems that already drive engineering and production.  ",{"type":25,"attrs":325,"content":326},{"textAlign":27},[327],{"text":328,"type":31},"That is especially important in the CAD-driven workflows common in manufacturing. If AI identifies a similar part, classifies a model, suggests a design change, or supports a manufacturing decision, engineers need access to the underlying CAD data and a way to evaluate the result in context. The system needs to preserve the connection between the AI output and the geometry, metadata, and engineering meaning behind it. ",{"type":25,"attrs":330,"content":331},{"textAlign":27},[332],{"text":333,"type":31},"For developers, the requirement is practical: AI workflows need to connect back to CAD, visualization, and data exchange infrastructure. Without that connection, AI risks becoming a disconnected layer rather than a useful part of the engineering process. ",{"type":25,"attrs":335,"content":336},{"textAlign":27},[337],{"text":338,"type":31},"This is where Tech Soft 3D’s broader toolkit portfolio becomes important. HOOPS Exchange supports CAD data access and translation across major engineering formats, while HOOPS AI supports the machine learning side of CAD workflows. Together, these capabilities help developers build AI-enabled applications that are grounded in real CAD data rather than isolated AI demos. ",{"type":110,"content":340},[341],{"type":25,"attrs":342,"content":344},{"textAlign":343},"center",[345,349,354,364,369],{"text":346,"type":31,"marks":347},"For teams thinking beyond CAD-based machine learning and into AI-assisted CAD generation, ",[348],{"type":293},{"text":350,"type":31,"marks":351},"Tech Soft 3D also resells ",[352,353],{"type":42},{"type":293},{"text":355,"type":31,"marks":356},"Parasolid",[357,362,363],{"type":94,"attrs":358},{"href":359,"uuid":360,"anchor":97,"target":98,"linktype":361},"/developers/products/parasolid","18051e61-383c-4411-bc10-fe7562f5a4fa","story",{"type":42},{"type":293},{"text":365,"type":31,"marks":366},",",[367,368],{"type":42},{"type":293},{"text":370,"type":31,"marks":371}," the world’s leading 3D geometric modeling kernel. Parasolid can support robust procedural CAD generation for applications that need to create editable, downstream-ready geometry.  ",[372],{"type":293},{"_uid":374,"height":375,"slides":376,"actions":460,"alignment":17,"component":461,"background_opacity":462},"8e58b363-73f0-4be7-b2c3-8c9c4523e485","height-compact",[377,408,436],{"_uid":378,"copy":379,"icon":386,"title":91,"action":17,"buttons":389,"eyebrow":17,"component":399,"icon_size":400,"background":401,"button_link":17,"button_label":17,"content_width":17,"title_heading":405,"button_new_tab":131,"background_video":406},"7ac0a2b9-2db7-42e5-a829-c202f83d10bb",{"type":22,"content":380},[381],{"type":25,"attrs":382,"content":383},{"textAlign":97},[384],{"text":385,"type":31},"Framework for CAD-based ML",{"id":97,"alt":97,"name":17,"focus":97,"title":97,"filename":97,"copyright":97,"fieldtype":387,"meta_data":388,"is_external_url":294},"asset",{},[390],{"_uid":391,"href":17,"color":17,"label":392,"action":393,"new_tab":294,"outline":294,"alignment":17,"component":394,"text_link":294,"destination":395},"c0940959-baa5-4fb1-a9dd-2a85b7127642","Explore HOOPS AI","action-link","element-button",{"id":396,"url":17,"linktype":361,"fieldtype":397,"cached_url":398},"d869c1a4-f588-4daa-8803-d688b86cee58","multilink","developers/products/hoops-ai","element-epic-child","large",{"id":402,"alt":17,"name":17,"focus":17,"title":17,"source":17,"filename":403,"copyright":17,"fieldtype":387,"meta_data":404,"is_external_url":294},193090599194239,"https://a-us.storyblok.com/f/1023536/5000x2813/3e93ceb759/adobestock_285865594.jpeg",{},"heading-h1",{"id":97,"alt":97,"name":17,"focus":97,"title":97,"source":97,"filename":17,"copyright":97,"fieldtype":387,"meta_data":407},{},{"_uid":409,"copy":410,"icon":417,"title":419,"action":17,"buttons":420,"eyebrow":17,"component":399,"icon_size":17,"background":427,"button_link":17,"button_label":433,"content_width":17,"title_heading":405,"button_new_tab":131,"background_video":434},"b80cd034-18d9-4aee-8f4b-3852dd7a58e6",{"type":22,"content":411},[412],{"type":25,"attrs":413,"content":414},{"textAlign":97},[415],{"text":416,"type":31},"The leading CAD data access library, with support for 30+ popular formats",{"id":97,"alt":97,"name":17,"focus":97,"title":97,"filename":97,"copyright":97,"fieldtype":387,"meta_data":418,"is_external_url":294},{},"HOOPS Exchange",[421],{"_uid":422,"href":17,"color":17,"label":423,"action":393,"new_tab":294,"outline":294,"alignment":17,"component":394,"text_link":294,"destination":424},"a5626f3b-569d-4172-881e-6718ed5fe882","Explore HOOPS Exchange",{"id":425,"url":17,"linktype":361,"fieldtype":397,"cached_url":426},"6bb4085d-b926-4852-81a8-7a02c45f5abc","developers/products/hoops-exchange",{"id":428,"alt":429,"name":17,"focus":17,"title":430,"source":17,"filename":431,"copyright":17,"fieldtype":387,"meta_data":432,"is_external_url":294},67798057063361,"A 3D grid of blue, transparent file icons","Data Exchange Formats","https://a-us.storyblok.com/f/1023536/1920x1427/273124f842/data-exchange-formats.jpg",{"alt":429,"title":430,"source":17,"copyright":17},"Explore HOOPS Envision",{"id":97,"alt":97,"name":17,"focus":97,"title":97,"source":97,"filename":17,"copyright":97,"fieldtype":387,"meta_data":435},{},{"_uid":437,"copy":438,"icon":445,"title":447,"action":17,"buttons":448,"eyebrow":17,"component":399,"icon_size":400,"background":454,"button_link":17,"button_label":17,"content_width":17,"title_heading":405,"button_new_tab":131,"background_video":458},"d155aa8f-64bd-4bcb-a23b-9b56c998e49c",{"type":22,"content":439},[440],{"type":25,"attrs":441,"content":442},{"textAlign":97},[443],{"text":444,"type":31},"Add advanced solids-based modeling to your 3D application with the leading modeling kernel, Parasolid.",{"id":97,"alt":97,"name":17,"focus":97,"title":97,"filename":97,"copyright":97,"fieldtype":387,"meta_data":446,"is_external_url":294},{},"Siemens Parasolid",[449],{"_uid":450,"href":17,"color":17,"label":451,"action":393,"new_tab":294,"outline":294,"alignment":17,"component":394,"text_link":294,"destination":452},"2678ddc7-4c7d-42f0-99e1-34335fd4e4af","Explore Parasolid",{"id":360,"url":17,"linktype":361,"fieldtype":397,"cached_url":453},"developers/products/parasolid",{"id":455,"alt":17,"name":17,"focus":17,"title":17,"source":17,"filename":456,"copyright":17,"fieldtype":387,"meta_data":457,"is_external_url":294},116319082020039,"https://a-us.storyblok.com/f/1023536/640x359/50efa37203/parasolid-ko2.png",{},{"id":97,"alt":97,"name":17,"focus":97,"title":97,"source":97,"filename":17,"copyright":97,"fieldtype":387,"meta_data":459},{},[],"element-epic","30",{"_uid":464,"body":465,"alignment":17,"component":130,"is_full_width":131},"03e0499e-152a-4d8d-a573-2900e660b87d",{"type":22,"content":466},[467,472,477,482,487,496,501,506,546,563],{"type":137,"attrs":468,"content":469},{"level":139,"textAlign":97},[470],{"text":471,"type":31},"Practical AI in Manufacturing Starts with the Right Data Infrastructure ",{"type":25,"attrs":473,"content":474},{"textAlign":27},[475],{"text":476,"type":31},"AI in manufacturing will not be defined only by model quality, but by whether AI systems can meaningfully work with the data manufacturing actually uses. ",{"type":25,"attrs":478,"content":479},{"textAlign":27},[480],{"text":481,"type":31},"That includes every part of CAD data, including B-Rep, engineering structure, human review, downstream compatibility, and the accumulated knowledge that makes each manufacturer’s products and processes unique. ",{"type":25,"attrs":483,"content":484},{"textAlign":27},[485],{"text":486,"type":31},"For software developers, these requirements are difficult to build from scratch. CAD access, geometry processing, visualization, machine learning preparation, and interoperability are all complex technical areas. Combining them into a reliable AI workflow can quickly become a major development bottleneck. ",{"type":25,"attrs":488,"content":489},{"textAlign":27},[490,494],{"text":491,"type":31,"marks":492},"HOOPS AI was built to help developers address that challenge",[493],{"type":42},{"text":495,"type":31},". Supporting machine learning workflows for CAD data, this toolkit gives teams a foundation for building practical AI capabilities into engineering and manufacturing applications. ",{"type":137,"attrs":497,"content":498},{"level":139,"textAlign":27},[499],{"text":500,"type":31},"Ready to Learn More about HOOPS AI? ",{"type":25,"attrs":502,"content":503},{"textAlign":27},[504],{"text":505,"type":31},"By far the best way to explore how HOOPS AI can help your organization apply machine learning to CAD data while protecting sensitive engineering information is to chat with us. ",{"type":25,"attrs":507,"content":508},{"textAlign":27},[509,516,518,525,527,533,535,539,544],{"text":510,"type":31,"marks":511},"Contact our team",[512],{"type":94,"attrs":513},{"href":514,"uuid":515,"anchor":97,"target":98,"linktype":361},"/contact","3e1a4870-930c-4660-9daf-91036e066d66",{"text":517,"type":31}," to learn how ",{"text":91,"type":31,"marks":519},[520,524],{"type":94,"attrs":521},{"href":522,"uuid":396,"anchor":97,"target":523,"linktype":361},"/developers/products/hoops-ai","_self",{"type":42},{"text":526,"type":31},", ",{"text":419,"type":31,"marks":528},[529,532],{"type":94,"attrs":530},{"href":531,"uuid":425,"anchor":97,"target":98,"linktype":361},"/developers/products/hoops-exchange",{"type":42},{"text":534,"type":31},", and",{"text":536,"type":31,"marks":537}," ",[538],{"type":42},{"text":355,"type":31,"marks":540},[541,543],{"type":94,"attrs":542},{"href":359,"uuid":360,"anchor":97,"target":98,"linktype":361},{"type":42},{"text":545,"type":31}," can fit into a broader engineering AI strategy. ",{"type":110,"content":547},[548],{"type":25,"attrs":549,"content":550},{"textAlign":97},[551,553,559],{"text":552,"type":31},"If you would like to explore how Tech Soft 3D’s toolkits can help you keep pace with a fast-moving industry, ",{"text":554,"type":31,"marks":555},"contact us to learn more",[556,558],{"type":94,"attrs":557},{"href":514,"uuid":515,"anchor":97,"target":98,"linktype":361},{"type":42},{"text":560,"type":31,"marks":561},".",[562],{"type":42},{"type":137,"attrs":564,"content":565},{"level":139,"textAlign":97},[566],{"text":567,"type":31},"FAQs About AI, CAD Data, and Manufacturing Applications ",{"id":569,"_uid":570,"items":571,"component":581},"desktop demo","8ea7d3ef-892d-4303-9ac8-a930957823c0",[572,574,579],{"_uid":573,"height":17,"component":18},"b0208994-9712-4667-b80a-69aa41f29c24",{"id":575,"_uid":576,"component":577,"business_unit":578},"c5a0aaa9-8ec9-4b18-a206-0d444fc65aa1","225620d9-8db5-43fb-950c-6ec32808bf09","hubspot-form","toolkits",{"_uid":580,"height":17,"component":18},"f437b9d5-79ea-4014-aba4-23f0880d5b1a","modal",{"id":583,"_uid":584,"items":585,"component":581},"talk to an expert","bf0c2043-997c-4416-a695-698515833a07",[586,588,591],{"_uid":587,"height":17,"component":18},"345c5aa5-84aa-4606-a8da-3abcda122edd",{"id":589,"_uid":590,"component":577,"business_unit":578},"169db079-f7dc-419c-b134-62ef63c9f571","b4686621-e8a4-414d-9198-eef6e8d07899",{"_uid":592,"height":17,"component":18},"ff41a3c7-240a-4fbe-829f-0a1c348bb829",{"id":594,"_uid":595,"items":596,"component":581},"web demo","989b0b9e-a75f-4264-9243-91aaff134f8a",[597,599,602],{"_uid":598,"height":17,"component":18},"41096758-7599-4412-b3de-5473483adf6a",{"id":600,"_uid":601,"component":577,"business_unit":578},"3fd4ef5c-6f92-4f7a-a8bc-44dccbf20f0a","0832c7b1-8582-42a2-9cf1-3498299b798e",{"_uid":603,"height":17,"component":18},"c09420ab-645a-437d-93ee-92ad3148c7b9",{"id":605,"_uid":606,"items":607,"component":581},"desktop analyzer","efc815e6-d30a-43cf-b247-ecbe75c5c03a",[608,610,613],{"_uid":609,"height":17,"component":18},"8b69358d-5fc3-4366-aa5b-bacf6fff0dae",{"id":611,"_uid":612,"component":577},"90925e0e-5bdc-4fb1-9f06-5a3c4826d793","e81e1ef1-6543-4d7e-8ffb-8a6a8ded2b80",{"_uid":614,"height":17,"component":18},"5196ec2f-f060-4110-bfdb-72db2329c25f",{"id":616,"_uid":617,"items":618,"component":581},"web analyzer","6fdec8cb-d54b-4ed7-8f18-a5d9bf3c2f88",[619,621,624],{"_uid":620,"height":17,"component":18},"9cd32293-a870-4720-a6eb-af3623f5ac33",{"id":622,"_uid":623,"component":577},"c4e2aa56-ee08-482d-9d91-579f29e0861e","f33ef774-6cb7-412e-bc35-cad2fe98810c",{"_uid":625,"height":17,"component":18},"4c475cea-efd8-4828-a25f-1d535ecc59db",{"id":627,"_uid":628,"items":629,"component":581},"desktop","744fb7a9-0491-4478-9ec9-675597942bae",[630,632],{"_uid":631,"height":17,"component":18},"8d354ba1-0d95-4963-bb68-057ec36927a1",{"id":633,"_uid":634,"component":577,"business_unit":578},"067ec949-b1d5-4f1b-920a-c163d9a4e777","f2945790-c68f-4c45-8a26-8dee13e4085e",{"id":636,"_uid":637,"items":638,"component":581},"web","8e203cf0-da82-4de7-96c7-4b9ed5da795d",[639,641],{"_uid":640,"height":17,"component":18},"7cac4801-5a25-4693-a66f-f8e3794478af",{"id":642,"_uid":643,"component":577,"business_unit":578},"cff84156-a91e-4a8d-96fd-362ed5f7618e","485622a1-4fdc-4284-9644-7bc565f21254",{"_uid":645,"height":17,"component":18},"70495671-d0ff-48be-8f5c-7576d1c00eeb",{"id":647,"_uid":648,"items":649,"component":581},"cta","0364f646-e409-4878-a346-35253e71b0bc",[650,652],{"_uid":651,"height":17,"component":18},"864121ae-1833-4267-a119-fd0817b8d8ef",{"id":589,"_uid":653,"component":577,"business_unit":578},"5a3ae18d-994f-4482-8c9d-1db9fa99a571",{"id":583,"_uid":655,"items":656,"component":581},"695e881d-e250-4d81-bd46-6bdb911ced1a",[657],{"id":658,"_uid":659,"component":577,"business_unit":17},"27c6a1c0-294d-4da5-9e1e-6069799a24c6","3ecc2abe-950d-411d-adf1-aeec496bc026",{"_uid":661,"title":662,"contents":663,"component":283,"description":664,"exclude_schema":294},"d9854620-cc57-419f-bf59-ef671f1be17a","Why is CAD data difficult for AI to use? ",[],{"type":22,"content":665},[666],{"type":25,"attrs":667,"content":668},{"textAlign":97},[669],{"text":670,"type":31},"CAD data is more complex than text or image data. A CAD file may contain BRep geometry, topology, dimensions, assembly structure, feature information, metadata, and other engineering context. For AI to use that data effectively, it needs to be read, interpreted, and converted into a machine-learning-ready representation. ",{"_uid":672,"title":673,"contents":674,"component":283,"description":675,"exclude_schema":294},"541de6d1-f961-4848-b47d-43abd95b30e6","Why can’t manufacturers just use general-purpose AI tools with CAD data? ",[],{"type":22,"content":676},[677],{"type":25,"attrs":678,"content":679},{"textAlign":97},[680],{"text":681,"type":31},"General-purpose AI tools are not usually built to understand raw CAD data directly. There is also a major data security issue: manufacturing CAD data is often proprietary and may contain sensitive product, customer, or process information. Many companies cannot upload that data to public or general-purpose AI services. ",{"_uid":683,"title":684,"contents":685,"component":283,"description":686,"exclude_schema":294},"46f7725d-f7c1-46c8-ae73-d2a8729bca1a","How does HOOPS AI help developers working with CAD data? ",[],{"type":22,"content":687},[688],{"type":25,"attrs":689,"content":690},{"textAlign":97},[691,696,699],{"text":91,"type":31,"marks":692},[693,695],{"type":94,"attrs":694},{"href":522,"uuid":396,"anchor":97,"target":98,"linktype":361},{"type":42},{"text":536,"type":31,"marks":697},[698],{"type":42},{"text":700,"type":31},"helps developers build machine learning workflows around CAD data. It supports reading CAD data, converting it into tensor-ready representations, and visualizing CAD models so users can evaluate machine learning results in context. This gives developers a CAD-focused foundation for adding AI capabilities to engineering and manufacturing applications. ",{"_uid":702,"title":703,"contents":704,"component":283,"description":705,"exclude_schema":294},"358f9599-c1fa-4008-91ef-1ce41fec230e","What questions should CAE teams consider when debating developing a feature internally?",[],{"type":22,"content":706},[707],{"type":25,"attrs":708,"content":709},{"textAlign":97},[710,715],{"text":419,"type":31,"marks":711},[712,714],{"type":94,"attrs":713},{"href":531,"uuid":425,"anchor":97,"target":98,"linktype":361},{"type":42},{"text":716,"type":31}," supports CAD data access and translation across major engineering formats. HOOPS AI supports machine learning workflows for CAD data. Used together, they can help developers build applications that access real CAD data, prepare it for AI workflows, and keep results connected to engineering data infrastructure. ",{"_uid":718,"messages":719,"component":742,"background":743,"remove_padding":131,"is_verticalized":294},"f9294fbe-31fc-414a-bcdd-813348c4dfee",[720],{"_uid":721,"icon":17,"buttons":722,"content":727,"headline":734,"link_url":17,"component":741,"icon_size":17,"link_label":17},"6720bf06-a4d9-499d-b738-7f5fdca52cc3",[723],{"_uid":724,"href":583,"color":17,"label":583,"action":725,"new_tab":294,"outline":294,"alignment":17,"component":394,"text_link":294,"destination":726},"d6e8d5bb-ce21-419d-9572-a96f64518f68","action-modal",{"id":17,"url":17,"linktype":361,"fieldtype":397,"cached_url":17},{"type":22,"content":728},[729],{"type":25,"attrs":730,"content":731},{"textAlign":97},[732],{"text":733,"type":31},"Chat with our team to see how Tech Soft 3D products could be right for you.",{"type":22,"content":735},[736],{"type":25,"attrs":737,"content":738},{"textAlign":97},[739],{"text":740,"type":31},"Want to learn more?","element-callout-child","element-callout",{"id":97,"alt":97,"name":17,"focus":97,"title":97,"source":97,"filename":17,"copyright":97,"fieldtype":387,"meta_data":744},{},"2026-06-30 00:00",[747,797],{"_uid":748,"source":749,"component":796},"d0505392-0fe2-4925-8e19-af7dcfee7b51",{"name":750,"created_at":751,"published_at":752,"updated_at":753,"id":754,"uuid":755,"content":756,"slug":787,"full_slug":788,"sort_by_date":97,"position":789,"tag_list":790,"is_startpage":294,"parent_id":791,"meta_data":97,"group_id":792,"first_published_at":793,"release_id":97,"lang":794,"path":97,"alternates":795,"default_full_slug":97,"translated_slugs":97,"_stopResolving":131},"Main Navigation","2025-04-09T14:04:01.876Z","2026-06-19T19:46:15.424Z","2026-06-19T19:46:15.434Z",27565407,"c5f4a23f-ca2c-41d7-946b-6822d34565a0",{"_uid":757,"items":758,"component":786},"8915070e-32eb-476d-b215-1c92ebdad15c",[759,768],{"_uid":760,"left_nav":17,"messages":761,"component":767,"right_nav":17},"fe7818a3-2c19-410c-adfd-6166e04c9bf0",[762],{"_uid":763,"icon":764,"link":514,"target":17,"message":765,"component":766},"87585227-383a-48e2-a5e5-1541d3b4b65a","tdesign:user-talk-1","Have questions about our SDKs or enterprise applications? Talk to a Tech Soft 3D expert","element-notice-child","element-notice",{"_uid":769,"menu":770,"component":785,"has_notice":131},"192904e7-cc6e-44e5-95a7-187c8ef72d1e",[771,776,779,782],{"url":772,"_uid":773,"label":774,"new_tab":294,"component":775},"#","eada60f3-c50f-41c1-8c50-4ffedd119483","SDKs for Developers","top-menu-child",{"url":772,"_uid":777,"label":778,"new_tab":294,"component":775,"style_as_button":294},"e80ecf91-c6c8-4e7a-a5b7-10ec4e05ada0","Apps for Businesses",{"url":772,"_uid":780,"label":781,"new_tab":294,"component":775,"style_as_button":294},"b3d091d9-83d1-4b72-bb8d-0f4be140beaa","Company",{"url":772,"_uid":783,"label":784,"new_tab":294,"component":775,"style_as_button":131},"f2040cd1-8c5b-4dcf-8ee1-5b56049d0fe2","Support","top","global","main-navigation","global/main-navigation",-30,[],27501948,"998e76b0-7b0e-4635-8abf-c9db0f0bf29c","2025-04-09T15:16:40.189Z","default",[],"reference",{"_uid":798,"height":17,"slides":799,"actions":816,"alignment":17,"component":461,"background_opacity":817},"993bae1d-856a-465b-b2be-11be3c50e671",[800],{"_uid":801,"copy":802,"icon":805,"title":807,"action":17,"eyebrow":808,"component":399,"icon_size":17,"background":809,"button_link":17,"button_label":17,"content_width":813,"title_heading":405,"button_new_tab":294,"background_video":814},"3e3045da-982d-442f-803c-c250ba2a89a2",{"type":22,"content":803},[804],{"type":25},{"id":97,"alt":97,"name":17,"focus":97,"title":97,"filename":97,"copyright":97,"fieldtype":387,"meta_data":806,"is_external_url":294},{},"Why CAD Data is the Hardest Part of Bringing AI to Manufacturing Applications ","Industry Insights",{"id":810,"alt":17,"name":17,"focus":17,"title":17,"source":17,"filename":811,"copyright":17,"fieldtype":387,"meta_data":812,"is_external_url":294},193090315824766,"https://a-us.storyblok.com/f/1023536/1900x600/c02521a64e/ai-in-manufacturing-hero.png",{},"width-medium",{"id":97,"alt":97,"name":17,"focus":97,"title":97,"source":97,"filename":17,"copyright":97,"fieldtype":387,"meta_data":815},{},[],"33","unit-toolkits",{"type":22,"content":820},[821],{"type":25,"attrs":822,"content":823},{"textAlign":97},[824],{"text":825,"type":31},"How do you integrate CAD data into AI workflows? In this blog, we explore what makes this issue so difficult in manufacturing workflows and the new technology that exists to finally solve the problem.","Koji Takaba",[828,830,868,870,880],{"_uid":829,"height":17,"component":18},"f3e70b03-6be1-4acd-a010-6c01b930e621",{"id":17,"_uid":831,"items":832,"component":867},"a217c55b-f796-41e6-b1eb-33c181b42d3c",[833],{"_uid":834,"colors":835,"padding":17,"contents":836,"component":858,"background_image":859,"background_video":865},"8a1ffd6e-0fdb-4ac0-9913-3a18762460b4","color-blue",[837],{"_uid":838,"icon":17,"image":839,"items":841,"title":842,"button":843,"eyebrow":17,"alignment":849,"component":850,"icon_size":17,"background":17,"description":851,"title_heading":17,"invert_text_color":131,"entire_region_linked":294},"da089f43-53ba-40a6-ad93-8941168b0595",{"id":97,"alt":97,"name":17,"focus":97,"title":97,"filename":97,"copyright":97,"fieldtype":387,"meta_data":840,"is_external_url":294},{},[],"Ready to Build Better Manufacturing Software?",[844],{"_uid":845,"href":17,"color":846,"label":847,"action":725,"new_tab":294,"outline":131,"alignment":17,"component":394,"text_link":294,"destination":848},"bfb4ada4-030a-4a0e-b532-250a56d33045","light","Talk to an Expert",{"id":17,"url":647,"linktype":99,"fieldtype":397,"cached_url":647},"align-left","element-feature",{"type":22,"content":852},[853],{"type":25,"attrs":854,"content":855},{"textAlign":97},[856],{"text":857,"type":31},"Learn how Tech Soft 3D Toolkits can give you a competitive advantage in developing manufacturing applications.","backdrop",{"id":860,"alt":861,"name":17,"focus":17,"title":862,"source":17,"filename":863,"copyright":17,"fieldtype":387,"meta_data":864,"is_external_url":294},108809046286679,"Abstract image of smooth, flowing white lines forming wave-like patterns on a dark blue gradient background.","Blue Wave Backdrop","https://a-us.storyblok.com/f/1023536/1920x1080/b8266f0798/2025-11-04_blue-wave-backdrop.jpg",{"alt":861,"title":862,"source":17,"copyright":17},{"id":97,"alt":97,"name":17,"focus":97,"title":97,"source":97,"filename":17,"copyright":97,"fieldtype":387,"meta_data":866},{},"group",{"_uid":869,"height":17,"component":18},"cca6faad-a2f5-49ef-a5f2-baff8e6cbf12",{"_uid":871,"body":872,"alignment":17,"component":130,"is_full_width":294},"4a86b1ad-d482-4436-839f-e19c37c78cc4",{"type":22,"content":873},[874],{"type":137,"attrs":875,"content":877},{"level":876},3,[878],{"text":879,"type":31},"Related Posts",{"_uid":881,"unit":818,"category":882,"component":884,"feed_type":885,"hide_filters":131,"vertical_feed":131,"items_per_page":17,"hide_pagination":131,"hide_descriptions":294,"hide_unit_selection":131},"24725ebb-5a13-4d0f-8e9c-430501195151",[883],"shipbuilding","feed","feed-blog",[887,888,889,890,891,892],"manufacturing","hoopsExchange","hoopsAi","parasolid","digitalConstruction","industrialDesign",{"_uid":894,"title":895,"plugin":896,"description":897},"f13492a7-770b-49c2-8089-8d76f977e868","Why CAD Data Is the Hardest Part of Bringing AI to Manufacturing Applications ","meta-fields","How do you integrate CAD data into AI workflows? In this blog, we explore what makes this issue so difficult in manufacturing workflows, and the new technology that exists to finally solve the problem.","post",{"id":900,"alt":17,"name":17,"focus":17,"title":17,"source":17,"filename":901,"copyright":17,"fieldtype":387,"meta_data":902,"is_external_url":294},193089652584057,"https://a-us.storyblok.com/f/1023536/620x349/508caeee8c/ai-in-manufacturing-thumbnail.png",{},{"id":900,"alt":17,"name":17,"focus":17,"title":17,"source":17,"filename":901,"copyright":17,"fieldtype":387,"meta_data":904,"is_external_url":294},{},"why-ai-in-manufacturing-needs-to-understand-cad-data","resources/blog/why-ai-in-manufacturing-needs-to-understand-cad-data","2026-06-30",-560,[],59097741247176,"ce324c48-1369-4bc1-94e5-ce0d2e94cbe1","2026-06-30T14:51:12.599Z",[],1782849628,[916],{"name":750,"created_at":751,"published_at":752,"updated_at":753,"id":754,"uuid":755,"content":917,"slug":787,"full_slug":788,"sort_by_date":97,"position":789,"tag_list":928,"is_startpage":294,"parent_id":791,"meta_data":97,"group_id":792,"first_published_at":793,"release_id":97,"lang":794,"path":97,"alternates":929,"default_full_slug":97,"translated_slugs":97},{"_uid":757,"items":918,"component":786},[919,922],{"_uid":760,"left_nav":17,"messages":920,"component":767,"right_nav":17},[921],{"_uid":763,"icon":764,"link":514,"target":17,"message":765,"component":766},{"_uid":769,"menu":923,"component":785,"has_notice":131},[924,925,926,927],{"url":772,"_uid":773,"label":774,"new_tab":294,"component":775},{"url":772,"_uid":777,"label":778,"new_tab":294,"component":775,"style_as_button":294},{"url":772,"_uid":780,"label":781,"new_tab":294,"component":775,"style_as_button":294},{"url":772,"_uid":783,"label":784,"new_tab":294,"component":775,"style_as_button":131},[],[],[],{"cache-control":932,"connection":933,"content-encoding":934,"content-type":935,"date":936,"etag":937,"referrer-policy":938,"sb-be-version":939,"server":940,"strict-transport-security":941,"transfer-encoding":942,"vary":943,"via":944,"x-amz-cf-id":945,"x-amz-cf-pop":946,"x-cache":947,"x-content-type-options":948,"x-frame-options":949,"x-permitted-cross-domain-policies":950,"x-request-id":951,"x-runtime":952,"x-xss-protection":953},"max-age=0, public, s-maxage=604800, stale-if-error=3600","keep-alive","gzip","application/json; charset=utf-8","Tue, 30 Jun 2026 20:01:17 GMT","W/\"81552a7115f8c7c6cb68b8ae4ec45eca\"","strict-origin-when-cross-origin","5.899.0","nginx/1.29.1","max-age=31536000","chunked","Origin,Accept-Encoding, Origin","1.1 3b2dd9ca40903562e8f0412d796f5600.cloudfront.net (CloudFront)","orwtZvqVePEMTzGV70R830GVnjbSl6OtDo1bWxNeWNzNHRsUeoWSXQ==","CMH68-P3","Miss from cloudfront","nosniff","SAMEORIGIN","none","e088f155-d856-464a-96f7-bca877e744e9","0.026237","0",1782849673609]