Agentic AI Meets 3D Geometry: What Actually Scales in Engineering (Metafold + Cosmon)

Описание: In this episode of AI Across the Product Lifecycle, Michael Finocchiaro is joined by Elissa Ross (CEO, Metafold) and Rui Aguiar (Founder & CEO, Cosmon) to unpack what’s real vs hype in engineering AI—especially the hard part: 3D geometry, manufacturing workflows, and adoption inside risk-averse enterprises. 

You’ll hear two contrasting approaches:
• Cosmon: AI-native, agentic workflow orchestration across CAD/simulation/PLM. 
• Metafold: geometry infrastructure—turning messy 3D representations into learnable, consistent features (so ML can actually work). 

Chapters / timings
00:08 Intro + guests (Metafold, Cosmon) 
00:33 Metafold: geometry → manufacturing decision intelligence 
01:26 Cosmon: agentic tooling across design → production 
02:41 “OpenAI moment” for engineering: why 3D ≠ text 
04:53 AI adoption lens + why UX mattered (chat interface) 
07:29 “One line in the middle of the screen” + product simplicity 
08:39 How startups use AI internally (coding, support, docs) 
10:05 Does AI change Agile vs Waterfall? (editors vs writers) 
12:31 Where users “encounter AI” in Cosmon vs Metafold 
14:28 Metafold’s “infrastructure” thesis for 3D learning 
17:00 Integrating into legacy CAD/PLM workflows (API-first) 
20:47 Skills for the next-gen engineer (first principles + AI) 
25:29 Will engineers need to be more multidisciplinary? 
28:34 Will there be an “OpenAI moment” for geometry? (skeptic take) 
29:19 What’s next after LLMs/MCP/agents—bullish vs skeptical 
33:30 Digital maturity: who can adopt engineering AI today? 
37:10 “Lightbulb moments” + change management reality 
40:03 Audience Q: IP protection + deployment options 
42:25 Parting thoughts: disruption, adoption curves, staying grounded 
43:56 Startup strategy: focus, survival, and narrow wedges 
45:53 Where to meet them: GTC, Hannover Messe, etc.