Ryan Stroupe, U.S. Army Corps of Engineers, and Chair-BCxA conference planning committee

At the 2025 BCxA Conference in Pittsburgh, the closing plenary featured an Oxford-style debate on the role of artificial intelligence in building commissioning (Cx). The topic of AI dominated the conference, and as the debate’s organizer and host, I offer my perspective on the industry’s current stance on AI, the debate itself, and the necessary next steps as the Building Commissioning Association navigates commissioning practices that will increasingly rely on AI.

Leading up to the event, I had a growing concern that our planned debate was scheduled two years too late. With AI tools heavily utilized by many BCxA members, I questioned if a debate on the topic would resonate if the proverbial ship had already set sail. This apprehension was cemented by my struggle to find speakers willing to voice a cautionary perspective on AI. To top it all off, the pro-AI debate team was more confident than Apollo Creed in the first Rocky movie. I feared the debate would be a blowout, but I held out hope for a lively and engaging discussion.

The debate's motion was “AI is making us better Cx providers.” The team arguing “For” the motion was comprised of John Villani of Grumman/Butkus Associates and Jesse Sycuro of McKinstry, while Theophilus Aluko of Willdan and Bill McMullen of Chinook Systems Inc. argued "Against." In this specific format, the winning team is not necessarily the one with the most supporters at the end of the debate. Instead, the winner is the team that was most persuasive and caused the biggest shift in audience opinion between the pre-debate and post-debate polling. Audience members could choose “For”, “Against”, or “Undecided” for each poll. Initially, 43% of the audience favored the motion, 17% opposed, and 40% were undecided, leaving the outcome still very much in doubt despite the pro-AI momentum.

Core Arguments

The Against Team opened by highlighting several risks:

  • The "Garbage In, Gospel Out” Dilemma: AI can produce outputs that appear authoritative but are dangerously misleading.
  • Erosion of Expertise and Trustworthiness: Over-reliance on AI could create professionals who can get an answer but cannot explain the “why” behind it.
  • Lack of Transparency: Many AI systems operate as "black boxes.”
  • Hidden Dangers: They pointed to privacy issues, data security concerns, intellectual property ownership, and liability risks.

The For Team countered with the benefits of AI:

  • Enhanced Capabilities: AI expands human limitations, allowing for faster, more accurate predictions.
  • Improved Detection: Machine learning can catch design and operational issues that humans might miss.
  • A Game Changer: They cited specific AI applications for Design Review, Functional Performance Testing (FPT), Energy Baseline & Optimization, Controls Verification, and Ongoing Commissioning (OCx).
  • Job Evolution: AI won’t take your job, but a professional using AI will.

Debate Highlights and Results

After the opening arguments, the discussion intensified with questions from the audience and opposing teams. Here are some highlights:

  • Theo Aluko opened with a memorable quip: 'Going to McDonald’s doesn’t make you a McChicken… AI doesn't automatically improve our work, arguing that AI makes us leveraged, not necessarily better Cx Providers.
  • Jesse Sycuro compared AI to 'the smartest intern possible'—a tool that does the heavy lifting but requires a professional to verify the work before handing it to a client.
  • An audience member asked the For Team, “What is the kWh price of each GPT query?” John Villani acknowledged that data center energy use is high but cited a research paper he had read (likely Theo Aluko’s ASHRAE paper) in which AI completed a two-day task in just 15 minutes. He argued that the cost savings from such efficiency far exceeded the energy investment.
  • Another audience member asked the Against Team, “Did you use AI to develop your slide presentation?” Theo Aluko admitted he used some AI. Bill McMullen did not, citing concerns about AI "hallucinations." The For Team seized on this, framing the issue as one of trust and comparing it to the medical industry's growing reliance on AI for reviewing MRI scans.
  • A question was posed to both teams: “Are you comfortable with turning in the working trail/audit log that AI creates to your client if they ask for it?” Jesse Sycuro of the For Team responded emphatically, “Absolutely!” He noted that the transparency of modern AI tools is a major improvement over the black boxes of the past. Bill McMullen from the Against Team countered that an AI audit trail is less valuable to clients than the established industry standards on which his firm relies.
  • Bill McMullen made a point about data security and uploading client-confidential information into an AI model, and the inherent risk of its loss. While John Villani argued that leveraging a Retrieval-Augmented Generation (RAG) methodology to search only your own, firewalled data would limit risk, it does not guarantee that customer-sensitive information remains indefinitely secure.
  • The audience reacted strongly at several points. Bill McMullen’s impassioned plea against the motion inspired a notable round of applause. Later, when he pointed out that the For Team’s own company websites didn't list AI as a core skill, the audience laughed and applauded.


In the end, the final poll showed 57% of attendees were in favor of the motion (a 14% increase), while just 20% were opposed (a 3% increase). The For Team was declared the winner. While it was a convincing victory, it is noteworthy that 23% of the audience remained undecided and that 9 of the 11 audience questions came from AI skeptics, which may suggest they, like me, were invested in producing a more competitive result. Even Jesse Sycuro of the For Team acknowledged that if given the opportunity, he would have voted for “Theo” and the Against Team.

Conclusion and Future Questions

It's clear that AI is already transforming our industry. However, the debate raised critical questions that the Building Commissioning Association must address:

  1. Should the BCxA create a “Cx with AI” best practice document covering liability, data privacy, and IP ownership?
  2. How can the industry promote Explainable AI (XAI) systems that can justify their reasoning?
  3. How should firms represent their expertise to clients — is it human knowledge, problem-solving capability, past experience, the ability to leverage AI, or a combination of these skills?
  4. How do we train future generations of Cx professionals who will have a tendency to leverage AI tools versus pursuing personal, first-hand experiences?
  5. How do we help owners distinguish between Cx firms that approach AI with quality control and transparency, and bad actors who use AI tools without quality control measures to reduce costs and underbid us on projects?

Given the level of interest and engagement on AI and Cx, I expect these concerns will fuel many conversations at upcoming BCxA leadership meetings and conferences. And I expect to see more debates like this as part of future BCxA conference agendas.