Continuing the AI Conversation
Last week’s deep dive into Kenny Seaton’s AI-driven transformation at CSU Dominguez Hills revealed how AI machine learning is revolutionizing building operations – from achieving 96% reductions in gas usage to creating self-optimizing systems that outperform static OEM sequences. This week, we explored the inevitable industry tension this creates: How can proprietary innovation coexist with the open ecosystems needed for AI’s full potential?
1. The Proprietary vs. Open Dilemma
The discussion opened with a fundamental question: Does innovation require proprietary technology? Anto framed the challenge using a light bulb (innovation) and padlock (proprietary) metaphor:
Key Observations:
- Historical Context: Early proprietary systems (like pre-BACnet controls) often emerged from necessity, not malice
- The Lock-In Effect: Some vendors later weaponized proprietary protocols to restrict customer choice
- Today’s Balance: “Proprietary innovation isn’t evil, but its implementation can be”
2. Industrial Automation’s Playbook
Highlighting the stark differences between building automation and industrial systems:
| Factor | Industrial Automation | Building Automation |
|---|---|---|
| Procurement | Equipment-driven (no low-bid) | Project-based (value-engineered) |
| Standards Adoption | Often viewed as a cost center | Proliferation of competing protocols |
| Mission Criticality | Direct revenue impact | Often viewed as cost center |
The Takeaway: Industrial sectors standardized because equipment failures resulted in immediate production losses—a lesson for buildings as they become revenue-generating assets.
3. The Platform Solution
The Linux Foundation’s Tiger Team emerged with a potential path forward:
Progress Report:
- RDF Framework: Consensus reached on using Resource Description Framework for cross-system data modelling
- GitHub Momentum: Code repositories now being populated with interoperable building data examples
- 2525 Integration: Work underway to align with ASHRAE’s building performance standards
“We’re not asking vendors to open their secret sauce – just the pantry doors. Let owners mix ingredients from different chefs.”
4. AI’s Double-Edged Sword
The group grappled with AI’s paradoxical role:
The Challenge:
- Proprietary algorithms deliver unprecedented optimization (like Kenny’s Chiller AI)
- Black-box models create new lock-in risks and auditability concerns.
The Opportunity: Keith’s case study showed how open APIs can allow:
- Third-party validation of AI decisions
- “Mix-and-match” with best-in-class analytics
- Future-proofing against vendor obsolescence
5. An Owner-Led Revolution
The session closed with clear action items:
For Owners:
- Specify RDF compliance in procurement docs
- Demand measurable outcomes in service contracts
- Support pilot projects, intelligence layers
For Vendors:
- Embrace “open enough” models while enabling integration
- Invest in developer ecosystems around your platforms
- Participate in standards groups like the Tiger Team
The Bottom Line:
As buildings evolve from static structures to adaptive organisms, the industry must forge a new compact one where proprietary innovation and open interoperability aren’t rivals, but essential partners in progress.
“Open (source) does NOT equal free”
I love this comment from one of my social media posts.
I also appreciate the balanced framing: No one is asking vendors to give away IP — but rather to allow data interoperability, validation, and modularity. This is essential for future-proofing both assets and the market itself.
Ken Sinclair This is exactly the kind of conversation we need to be having — not just about what AI can do for buildings, but how we govern the innovation ecosystem that enables it.
Kenny’s example at CSU Dominguez Hills is a blueprint for the next-gen built environment: AI-powered, performance-driven, and sustainability-first. But as highlighted, the true complexity lies in the ownership and interoperability of intelligence. Proprietary optimization engines offer near-magical efficiencies, yet often at the cost of flexibility, transparency, and long-term integration capacity.
🔍 What stood out most:
• The light bulb vs. padlock metaphor captures the conflict beautifully: Innovation isn’t the enemy — enclosure is.
• Drawing lessons from industrial automation’s maturity is crucial. Building automation still suffers from fragmented procurement logic and cost-center mentality.
• The Tiger Team’s use of RDF and open GitHub frameworks could finally deliver a scalable foundation for semantic interoperability — something the smart buildings sector desperately lacks.
I also appreciate the balanced framing: No one is asking vendors to give away IP — but rather to allow data interoperability, validation, and modularity. This is essential for future-proofing both assets and the market itself.
📌 A few reflections:
• Owners should absolutely specify RDF/2525 compliance and AI validation layers as baseline requirements. Procurement drives transformation.
• Vendors embracing “open enough” strategies — think open APIs + closed models — may dominate the next 5 years.
• The biggest missed opportunity in many projects isn’t a lack of smart tech, but a lack of architecture for continuous learning and adaptability.
https://www.bimhero.io/posts/86656765/comments/133148154?utm_source=manual
Watch the Full Discussion:
Continue the Conversation: #SmartBuildings #OpenInnovation #AIforBuildings
AutomatedBuildings.com – Where the smart building industry builds its future.
Miss last week’s Monday Live?
