Beyond Automation: Orchestrating Moldflow with AI Agents

Extending Synera's agents beyond automation: scripting and gate optimization

Over 50% of engineering time goes into handoffs, file transfers, and switching between tools rather than the engineering work itself. That's the gap agentic AI for engineering is built to close, and recorded at CONNECT, the European Moldflow user meeting, this session shows Hendrik Schütte, Forward Deployed Engineer at Synera, taking it further. A supervisor agent routes each request from one chat window to specialist agents that collaborate on the task: a manufacturability agent checks wall thickness on an uploaded CAD part, a de-featuring agent removes a leftover tool mark, and the system rotates the part on its own to clear a draft-angle problem before the Moldflow agent runs a best-gate analysis and a full filling simulation. No file ever leaves the chat.

The session then goes beyond that core workflow. It shows a scripting agent that writes and runs a Moldflow reporting script on request, and a gate-optimization agent that tests several candidate injection points along a part's centerline and returns comparative simulation results for review. NASA used the same agent-based approach to go from two engineers exploring four design variants over two days to one engineer exploring more than 100 variants in an hour.

Key takeaways

• How a supervisor agent routes requests to specialist agents for geometry cleanup, manufacturability, and Moldflow analysis, all collaborating in one chat window
• How the system removes a leftover tool mark from a part and rotates it automatically to resolve a draft-angle problem, without opening a CAD tool
• How to run a Moldflow best-gate analysis and a full filling simulation from chat, then move the gate location and re-run without touching the file
• How a scripting agent that knows the Moldflow documentation writes and runs a script on request, for one-off reporting tasks like exporting specific results to a summary
• How a gate-location-optimization agent tests multiple candidate injection points along a part's centerline, decides for the variant with best balancing and returns comparative simulation results for review 
• What a built-in check looks like when the system flags a missing material input before a simulation can run

Who is this recording for?

• Injection molding and simulation engineers who want to cut time spent on manual manufacturability assessment and design or tooling optimization
• Simulation team leads deciding where collaborating specialist agents can extend past single-task automation into a full manufacturability-to-simulation workflow
• Digital transformation leads at OEMs and manufacturers evaluating agentic AI for engineering across CAx and simulation tools