AHR Expo 2026 | Education Session Recap
Session Recap
Most building owners believe their efficiency work is done. They have upgraded equipment, implemented sequences, hit a code benchmark, and moved on. This AHR Expo 2026 session challenged that assumption directly and made a compelling case that the efficiency gap is far wider than the industry tends to admit.
The panel brought together five practitioners working at different layers of the problem: building automation and fault detection, domestic hot water and CO2 heat pump systems, AI-powered central plant optimization, and federal and state tax incentive programs.
The Efficiency Gap Is Real and Largely Invisible
The session opened with a pointed observation: efficiency is not a one-time event. Buildings that were carefully tuned six weeks ago may already have had every programmed sequence manually overridden by operators managing comfort complaints across dozens of sites. No savings are being captured. No one has flagged it. The building appears to be running fine.
This is the efficiency gap. The distance between what a building is theoretically capable of and what it actually delivers day to day.
Field data from the building automation side was striking. Roughly 90% of commercial buildings carry a 25 to 30% HVAC savings opportunity simply from better optimization of existing equipment, approximately 13% of total building energy use. Guideline 36 has raised the baseline for control sequences, but those sequences are still hard-coded into controllers that do not learn, do not adapt, and do not catch problems when sensors drift or actuators fail.
AI: What It Actually Does in Buildings
A persistent misconception was addressed directly. AI in buildings is not a chatbot. It does not need to communicate in language. It speaks in numbers, and its job is to communicate with equipment, not with people.
The relevant comparison is not a large language model. It is a system that monitors every data point across an entire central plant every few minutes, makes micro-adjustments based on real-time conditions, and keeps learning from the outcomes. No human operator, however skilled, can match that combination of speed, scope, and continuity.
The hallucination concern that often comes up around AI in critical systems was also addressed. Not all AI operates the same way. Physics-based, iterative learning models of the type used in central plant optimization build from observed baseline behavior and make bounded adjustments within known operating ranges. Hallucination is a property of large language models parsing ambiguous human input. It simply does not apply in the same way. The practical advice: ask vendors directly what type of AI they are using, how it learns, and what limits are placed on its decisions.
A case study brought the numbers into focus. Cal State University Dominguez Hills installed what was described as the largest heat pump installation west of the Mississippi. Their energy bill was expected to rise. Instead, it fell by $250,000 as a direct result of AI optimization layered on top of the new equipment.
Fault Detection: The Necessary Foundation
A clear line was drawn between optimization and fault detection, and why both are needed. Optimization keeps a properly functioning system running as efficiently as possible. Fault detection catches the moment something stops working: a failed sensor, a stuck actuator, a drifting reading that quietly undermines every sequence downstream.
A CO2 sensor reading zero in an unoccupied space. A supply temperature that looks wrong. A condenser water set point that has been overridden and forgotten. These are the conditions that make optimization sequences irrelevant. Fault detection finds them.
The recommendation for building owners with a BAS but no current analytics layer: start with fault detection, then layer in optimization. The two tools are complementary, not competing. Roughly 30 established companies now offer FDD platforms, with AI-based optimization products newer but growing quickly.
From the smaller-scale side, a practical counterpoint emerged. For apartment buildings running domestic hot water heat pump systems, often without any BAS at all, the monitoring gap is not sophisticated. It is basic. Equipment fails with error codes that nobody reads. The first sign of a problem is a cold shower. The solution at this level is simple controller logic that monitors amperage, detects an enabled unit drawing no current, and sends a text to the site supervisor. Not AI. Not FDD. Just a signal that something needs attention.
The principle is the same at every scale: you cannot manage what you cannot see.
Value Stacking: Efficiency and Decarbonization Together
Field experience consistently shows that efficiency and low-carbon goals reinforce each other when approached together rather than in sequence. In one installation, a domestic hot water CO2 heat pump unit was repositioned inside the building’s games room. The heat rejection from producing hot water cooled the space throughout summer at no additional capital or operating cost. The heat pump manufacturer now features the approach in their own materials.
The economics of low-carbon upgrades, including CO2 heat pumps, solar thermal, and heat recovery systems, improve substantially when efficiency measures are woven into the same project. Payback periods shorten. Utility incentives stack. Occupant comfort benefits become part of the business case.
Tax Incentives: Underutilized at Every Level
The tax incentives portion of the session covered ground that reliably generates questions at every AHR appearance. The legislative environment has shifted. The One Big Beautiful Bill changed elements of what the Inflation Reduction Act had established, but the common assumption that federal incentives have largely disappeared is incorrect.
The 179D deduction has existed for 20 years. It applies to energy efficient commercial building improvements and can be allocated from tax-exempt building owners, such as government buildings and nonprofits, to the contractors and engineers who designed the qualifying systems. Many contractors who have successfully claimed it still encounter colleagues hearing about it for the first time.
The practical guidance: incentives at the federal, state, and utility level can be stacked on a single project. Utility programs are the most familiar entry point, but they frequently coexist with state or city programs and federal deductions that go unclaimed. Projects completed two or three years ago that triggered incentives are still being discovered today.
For project teams, timing matters. Bringing a tax incentives specialist in early, at the design stage rather than after construction, enables better qualification, higher claim values, and projects structured to hit multiple program thresholds simultaneously.
One strategy that has gained traction with large organizations pursuing long-term sustainability commitments: coordinate across accounting and facilities teams to create a formal link between projected incentive values and the annual facilities budget. If future incentives are reasonably predictable, they can justify a higher budget today, funding more projects, generating more incentives, and creating a self-reinforcing cycle.
Operators: The Most Underestimated Variable
Every topic in the session eventually circled back to operators. The consensus was consistent: technology alone does not hold. If the people managing a building do not understand, trust, or have time for an efficiency system, the system will be overridden. Often within weeks.
This is not a criticism of operators. Projects are frequently sold at an executive level and pushed down to facility staff who had no involvement in the decision. Operators are managing dozens of buildings, fielding comfort complaints, and doing so with limited time. Complexity introduced from above without adequate training or support will be bypassed.
The solutions that work are concrete. Dashboards that surface energy KPIs alongside comfort data so operators see both at once. Performance-linked incentives for facility staff. Clear, concise operator guides rather than dense technical documents. Involving operators in the design and commissioning process rather than presenting them with a finished system to maintain.
A longer-term perspective was also raised. The generation now entering the workforce has grown up with AI tools and treats them as standard problem-solving infrastructure. The cultural friction between operators and automated systems is likely to ease as that transition progresses.
The practical first step toward operator buy-in remains simple: reduce their load. Show them that the new system handles things they were previously doing manually. Create capacity before asking for engagement. The operator who has more time is the operator who has room to care.
Key Takeaways
- Buildings that appear efficient often are not. Scheduling overrides, failed sensors, unmonitored equipment, and deferred maintenance quietly erode savings that nobody is tracking.
- Continuous monitoring, whether through fault detection, AI optimization, or both, is how the gap gets closed. It is not a one-time project. It is an ongoing operational practice.
- AI in building systems is physics-based and bounded, not the same as language models. It communicates with equipment, not people, and it does not take vacations or miss the 3 AM shift.
- Incentives exist at the federal, state, and utility level and can be stacked on a single project. Many go unclaimed. Engaging specialists early improves outcomes at every level.
- Operator engagement is not optional. Systems that operators do not trust will be overridden. Reducing complexity and workload first creates the conditions for genuine buy-in.
- Efficiency and decarbonization are stronger together. The business case for low-carbon retrofits improves significantly when efficiency measures are part of the same project scope.
Speakers
Stephanie Poole, Principal, SES Consulting (Moderator) LinkedIn
Abby Massey, Principal, Technical Services & Business Development, Capstan Tax Strategies LinkedIn
Danielle Radden, VP of Global Revenue Operations, Facil.ai LinkedIn
Terry Herr, Principal, Intellimation LLC LinkedIn
Scott Graham, Owner, Renew Energy LinkedIn
