Key Takeaways:
Artificial intelligence (AI) is rapidly changing the way commissioning professionals work — but perhaps not in the way many people expect.
Building analytics platforms have been detecting operational faults such as simultaneous heating and cooling, short cycling equipment, sensor drift, and scheduling issues for years using deterministic, rule-based logic. These tools remain extremely effective at identifying that something isn't operating correctly.
The greater challenge has never been finding faults. It's determiningwhythey occurred.
Modern commissioning projects generate enormous amounts of information — from building automation system (BAS) trend data and functional performance tests to sequences of operation, owner project requirements (OPR), equipment submittals, issue logs, and design documents. While analytics can identify operational anomalies, the information that defines what should happen often lives in project documents that were never created to be machine-readable. Engineers have traditionally spent significant time connecting detected faults to design intent, project history and the most likely root cause.
The real shift is not that AI can detect faults better than existing analytics platforms. It is that large language models (LLMs) can help engineers connect those faults to the project documentation that defines how systems were intended to operate — turning sequences of operation, OPRs, submittals, issue logs, and other historically hard-to-search information into usable context for root-cause analysis.
Rather than replacing engineering judgment, AI helps commissioning teams process and understand large volumes of technical information, allowing engineers to focus more of their time on solving problems instead of searching for them.
Traditional building analytics answer an important question: What happened?
AI helps answer the next question: Why did it happen?
Consider an economizer that consistently fails to operate when outdoor conditions are favorable. Rule-based analytics can detect that the dampers remain closed while mechanical cooling continues to operate. What those rules cannot determine is whether the problem stems from an incorrect sequence of operation, a programming error, a failed actuator, a sensor calibration issue or an intentional override made during construction.
Answering those questions has traditionally required engineers to review multiple sources of project documentation, including sequences of operation, control drawings, equipment submittals, issue logs and previous commissioning reports.
LLMs dramatically reduce the time required to search, compare, and synthesize this information. By helping engineers connect operational faults with the documents that define how systems were intended to operate, AI accelerates root-cause investigation and prioritization while leaving engineering judgment exactly where it belongs—with experienced commissioning professionals.
Another significant shift is occurring behind the scenes. Historically, commissioning teams often adapted their workflows to match the limitations of commercially available software platforms. Recent advances in AI have dramatically lowered the cost and complexity of developing purpose-built tools that support the way commissioning professionals actually work.
At EEI Building Performance, this shift is reflected in the development of the EEI Data Gateway, which streamlines the collection, organization, and preparation of building performance data for commissioning activities. Collecting and normalizing data across different building automation systems has historically been a painful, high-cost lift — the kind of problem a firm our size would typically solve by buying a platform and adapting to it. What changed is the economics of building it ourselves. AI compressed the development effort enough that EEI could design the Gateway around the exact workflow its engineers needed, rather than waiting on a vendor's product roadmap. When new requirements emerge, they get built.
Similarly, digital commissioning platforms such asBalanceCxcentralize functional testing, deficiency tracking, project documentation, and communication into a single environment. AI enhances these workflows by organizing information, summarizing project documentation, and helping engineers quickly locate the information needed to investigate deficiencies.
The result isn't automated commissioning — it's more efficient commissioning.
Most commissioning projects generate far more information than engineers can realistically review in detail.
Thousands of BAS trend points, hundreds of equipment submittals, sequences of operation, test reports, RFIs, issue logs, and field observations all contribute valuable information, but reviewing each source manually is both time consuming and expensive.
AI helps commissioning teams prioritize where engineering effort delivers the greatest value. Instead of replacing existing analytics, AI complements them by organizing technical information, highlighting relationships between documents, surfacing probable root causes, and identifying the issues most deserving of engineering review.
Detection was never the industry's primary bottleneck. Engineering capacity was.
By reducing the time required to locate information, compare documentation, and investigate probable root causes, AI allows experienced commissioning professionals to focus on technical decisions that improve building performance.
AI is an extraordinarily capable research assistant, but it is not a commissioning authority.
AI cannot evaluate whether the owner's operational goals have changed, recognize constructability issues observed in the field, balance competing operational priorities, or determine the most appropriate corrective action. Those decisions require engineering experience, collaboration with project stakeholders, and an understanding of how buildings actually operate.
The future of commissioning is not about replacing engineers with AI. It is about giving experienced engineers better tools to process information, investigate problems more efficiently and make higher-quality decisions.
As AI continues to mature, its greatest impact on commissioning will likely come from augmenting engineering expertise rather than automating engineering decisions.
The firms that create the most value will not simply adopt AI tools — they will integrate AI into proven commissioning processes, enabling engineers to work more efficiently without compromising technical rigor.
At EEI, we see AI as another step in the evolution of commissioning. Rule-based analytics remain invaluable for detecting operational faults. AI extends those capabilities by helping engineers understand the engineering context surrounding those faults, prioritize corrective actions, and manage increasingly complex projects with greater efficiency.
For owners, the result is better-informed decisions, improved building performance, and commissioning teams that can deliver greater value throughout the life of the facility.
Joshua Keeler is EEI’s senior software engineer and lead developer with over 15 years of experience in building automation and computer science. He works closely with engineers and technicians to customize and advance BalanceCx, EEI’s data analytics software platform. Keeler specializes in troubleshooting complex automation challenges and developing reliable software architectures and APIs. Through active collaboration with operations staff and project stakeholders, he aligns software integrations with each owner’s specific needs. His work supports EEI’s delivery of innovative, high-performing building automation systems that strengthen building operations and performance.
