If AI was the buzzword of 2024, then 2025 is proving to be the year that the building automation and smart systems community began to move from hype to application. This year’s Haystack Connect conference confirmed a clear shift: from conceptual digital twin aspirations toward the tangible promise of automated tagging, ontology integration, and the real use of AI and large language models (LLMs) to improve engineering workflows.

Dr. Shadi Tabasi and Chris Hartley, members of Altura’s analytics team—joined the community to see where the latest trends intersect with the solutions and services we provide you in building analytics, master systems integration, and ontology development. Here’s what we saw—and what you should know if you’re in charge of managing building performance, deploying smart systems, or working to future-proof your facilities.

Major Themes from the Haystack Community

1. AI has gone from just ‘talk’ to implementation—it’s showing up in the toolchain

From converting PDFs to structured JSON models to streamlining data model generation, AI is increasingly being used to reduce time spent on repetitive, manual tagging. LLMs are being trained on “intelligence capital”—human expertise codified into reusable processes—to bridge the growing labor gap in facilities management and commissioning.

Why it matters: For facility managers, engineers, and integrators, AI tools can drastically reduce the time it takes to get from raw data to swift and actionable insight—without waiting for overburdened human teams to keep up. (See our latest UNR post on speed to connectivity!)

2. Digital twins have taken a back seat

Surprisingly, “digital twin” barely registered at this year’s event—a marked shift from past years where it dominated discussion on best practices. Attendees are clearly looking for more practical, sustainable approaches to modeling that don’t require constant re-syncing and unrealistic maintenance.

Why it matters: The market is realizing that long-term models are only useful if they can be maintained easily. Our use of SkySpark and feature-driven M&V (Measurement & Verification) is proving more reliable and useful than full-scale “twin” implementations for many clients.

3. Xeto and Haystack 5: The next evolution in data modeling

The conversation around ontologies and semantic tagging has matured. Just last year, Shadi’s RDF presentation was a pioneering entry point—this year, ontology frameworks like SHACL and RDF were widely understood and discussed.

Why it matters: As contributors to Xeto (the open-source tagging ontology evolving from Project Haystack), Altura is actively helping shape the future standard. This gives our clients a long-term edge in flexible, interoperable data models that aren’t locked into one vendor ecosystem.

4. Vendor lock-in and tagging still create friction

The conference reiterated an industry-wide challenge: inconsistent or nonexistent tagging at the point of device installation. Without early tag implementation, integrators are left scrambling to interpret systems post-installation—adding days of manual effort.

Why it matters: Altura’s approach continues to beat industry averages on integration speed because of our in-house tools, trained team, and forward-thinking use of AI. We’re focused on empowering engineers to do more engineering—less data wrangling.

5. Aging workforce + institutional knowledge loss = AI opportunity

The industry is facing a major turnover in institutional knowledge. Conference attendees widely acknowledged that many systems were built and maintained by a generation now approaching retirement.

Why it matters: Capturing, structuring, and embedding this expertise into AI-enhanced workflows helps prevent critical knowledge from walking out the door. Altura is actively building this into how we support ongoing operations for clients.

💡 BONUS: Advancing Semantic Interoperability – Shadi Tabasi’s Presentation on RDF & Xeto

One of the highlights of the conference was Altura’s own Shadi Tabasi’s presentation: “Haystack’s Pathway to RDF: Advancing Semantic Interoperability with Xeto and RDF”.

Her session broke down the technical transition from traditional tagging methods toward more scalable, machine-readable ontologies. By integrating Resource Description Framework (RDF) and tools like Xeto, her work positions Haystack’s evolution toward broader interoperability with other frameworks like ASHRAE 223P, Google Digital Buildings, and Brick Schema.

Why it matters: RDF-based tagging supports real-time validation, easier data sharing, and a unified way to describe complex building systems—critical to scaling smart building infrastructure without relying on proprietary lock-ins.

For those who missed it, the key message was clear: the future of building data is not only tagged, but semantically rich and interoperable by design—and Altura is helping pave the way.

What Now? How to Take Action

As a facilities leader, building owner, or sustainability manager, here’s what you can do with this information:

  • Ask your integrator about semantic tagging early in your project lifecycle—preferably at BIM stage or spec level. Altura can help standardize and validate these tags using RDF and SHACL.
  • Evaluate your existing data infrastructure—Are you still manually tagging devices post-installation? Let’s talk about how AI can help streamline that.
  • Audit your vendor lock-in risks—Flexible, vendor-agnostic ontologies like Xeto can make your smart building investments resilient over time..
  • Explore our SkySpark-based M&V workflows—We’ll show you how we deliver faster insights without waiting on “digital twins” to be updated.

Final Takeaway

The 2025 Haystack Connect conference confirmed what we’ve been building toward: smarter systems start with structured data, and the future of that structure lies in open ontologies, intelligent automation, and a proactive approach to capturing expert knowledge. With Xeto and Haystack 5 on the horizon, and AI tools becoming more embedded in the engineering workflow, Altura remains at the forefront—making complex systems simpler, more scalable, and more sustainable for our clients.

If you’re interested in finding out more about how Altura and our team are ensuring optimal building performance with clean data structures and ontology best practices, reach out to us, or feel free to drop Shadi (STabasi@alturaassociates.com) or Chris (chartley@alturaassociates.com) a line to chat directly about these industry impacts shaping your building data.