AISB Research · original, source-verified reportsSubscribe →
● AISB RESEARCH · UPDATED JUNE 2026

Research

AISB Research is original, source-verified work on how commercial buildings actually run — and, increasingly, how AI agents read them. No vendor briefings, no demos: every claim is checked against the public record, and what we can't verify, we say so.

Featured Series
The Agent-Native Interoperability Series
The agent is the adapter; the law is the lever; the benchmark is the moat.
6 reports · 1 interactive tool · open data
The Reports
№ 01FLAGSHIP
The APAC Building Data Interoperability Report 2026
To our knowledge the first interoperability assessment written for Asia-Pacific. Certification-rich, interoperability-poor — why NABERS, Green Mark, EEWH and CASBEE sit outside the platforms owners actually buy.
Read →
№ 02GLOBAL REPORT
The State of Building Data Interoperability 2026
A building can be 100% BACnet-compliant and 0% understandable. The standards reality check, the middleware decade, and why the next attempt won't look like middleware.
Read →
№ 03REPORT CARD
The Building Data Openness Report Card v1
Twelve platforms scored on data openness from public evidence only — 120 points possible, 8 awarded. Every grade citation-traced; every vendor a right of reply.
Read →
№ 04BENCHMARK
We Benchmarked LLMs on the Ugliest Problem in Building Data
Frontier models 80.8% exact, fast tier 48.3%, on raw BMS point mapping. 265 points, open scorer, CC-BY corpus. The number is now public.
Read →
№ 05OPEN SOURCE
Three Open MCP Server Templates for Building Data
10,000+ MCP servers exist; we found one for buildings. Green Button, Haystack, Brick — read-only by design. The building-side blanks, filled.
Read →
№ 06COMMUNITY DRAFT
The Agent-Ready Building Checklist v0.1
14 questions decide whether an AI agent can read your building. A community draft, not a standard — most buildings fail on the contract clause, not the technology.
Read →
The Tool
Interactive · Advisory-only
The Agent-Ready Building Scorecard
Score your building in 14 questions across 4 dimensions — Data Export, Semantic Readiness, Access Governance, Agent Surface. Nothing here touches your BMS.
Score your building →
Open Data
The benchmark corpus (265 points, 12 sets) is published CC-BY-4.0; the three MCP server templates are open and read-only by design. While the public repository is being set up, request either from hello@ai-smart-buildings.com — contributors are credited in v0.2.

Compiled by the AISB agent fleet from primary sources; every claim verified against the public record. Cost figures are labeled industry estimates. Source lists available on request.

✉️ The Intelligent Building Brief — get each report as it lands · 🤖 Ask our agents — free CRE analysis