Manufacturers have long envied the automation breakthroughs of the software world, but in design engineering, decades of critical know-how remain trapped in unstructured formats. Daikin Industries is taking a bold step to change that, selecting Outerport as its core platform to extract structured data from technical drawings, diagrams, and other complex engineering documents.
The AI bottleneck in manufacturing
While AI agents like Claude Code are revolutionizing software development, manufacturing has lagged. Legacy design assets—PDFs, diagrams, schematics, and even paper documents—aren’t easily digestible by AI. Without clean, structured data, even the smartest models can’t automate critical design processes.
Outerport aims to close that gap. Using proprietary computer vision and multi-modal large-language models (MLLMs), Outerport converts unstructured visual data into machine-readable formats like JSON. The result: AI agents can analyze, reason, and act on decades of engineering IP, enabling automation of tasks previously impossible to scale.
Why Daikin chose Outerport
Built by former research engineers from NVIDIA and Meta Platforms, Outerport impressed Daikin with both extraction accuracy and compliance with stringent information security standards. Shohei Hido, Executive Engineer at Daikin’s Technology Innovation Center, praised the platform:
“We had conducted internal trials on our technical documents but struggled to achieve the accuracy required for business operations. Outerport provided a dramatic leap in parsing precision and, importantly, offered agile customization to meet our specific needs.”
The deployment of Outerport is more than a technical upgrade—it’s the foundation for a future AI-driven engineering stack at Daikin, allowing the company to scale AI agents across its design organization and automate high-value engineering workflows.
Building an AI-first engineering workflow
Outerport’s vision centers on accelerating hardware development by catching issues early in the design phase with AI-driven checks. Its platform focuses on three pillars:
- Document extraction: Turning complex engineering diagrams into LLM-ready formats.
- AI Agent Systems: Empowering manufacturers to build in-house AI agents that leverage structured data.
- Domain expertise: Applying specialized knowledge for engineering and manufacturing processes.
By combining these elements, Outerport hopes to help manufacturers like Daikin shorten development cycles, reduce errors, and unlock new efficiencies in design engineering.
For manufacturers, the lesson is clear: automation in software is easy; automation in engineering is hard. Outerport and Daikin are demonstrating that with the right tools, even the most complex design workflows can finally join the AI revolution.
Power Tomorrow’s Intelligence — Build It with TechEdgeAI
