Temporary building overrides need an expiry plan
Research perspective | Logic Foundry | October 9, 2026. This is a proposed workflow for evaluation, not a report of a completed Logic Foundry experiment or an available autonomous control feature.
1. The problem is lost intent
Imagine a fictional office building that extends an air-handling schedule for an evening event. The adjustment is appropriate that night. Weeks later, the extended hours remain, but the reason is no longer visible to the operations team. A dashboard may show unexpected operation without explaining whether it is a mistake or a continuing requirement.
The useful question is whether the current behaviour still has an authorised purpose. An exception with context is different from an unexplained deviation. That distinction matters when designing future sensor analytics and maintenance workflows.
2. What published guidance supports
The U.S. Department of Energy Federal Energy Management Program describes re-tuning as a process for finding and correcting operational problems using building automation data. Its guidance includes collecting building information, analysing trends, walking the building, implementing appropriate actions, reporting findings and continuing the practice through operations and maintenance. This is U.S. federal-building guidance, accessed October 9, 2026, rather than validation of our proposed workflow or a universal requirement. [1]
Our inference is that operational continuity deserves attention alongside fault detection. The guidance supports an ongoing process; it does not establish that the exception record proposed below improves outcomes. We make no energy-savings estimate for that proposal.
3. Design proposal: give each exception a review point
A temporary override could carry a compact record: the equipment and point, the authorised reason, the responsible person, the intended duration and the evidence needed at review. Preserve the previous intended state and any constraints the operator must check before changing it.
An expiry plan should mean a planned review, not automatic reversal in every situation. Occupancy may have changed or work may still be underway. A qualified operator should confirm the appropriate action under existing site procedures, especially where ventilation, freeze protection or other protective functions could be affected.
A future interface could separate exceptions into three states: justified and within the review period; due for review; and unresolved because evidence or responsibility is missing. This is a proposed information design, not a shipped product function. The purpose would be to help people prioritise, rather than declare every exception a fault.
4. Where the idea can fail
A neat record is not proof that the underlying point is correct. Incorrect asset mapping, stale timestamps, disconnected sensors and undocumented local controls can undermine interpretation. The workflow also fails if the named owner lacks authority, or repeated extensions conceal a permanent change that should be documented properly.
Excessive administration creates another failure. If recording an authorised change takes longer than the task itself, people may stop maintaining the record. A trial should test both decision quality and the burden of capturing context.
5. A small evaluation a facilities team could try
For a proposed four-week observational trial, choose one building and a small set of non-critical scheduling exceptions. Keep existing control authority and safety procedures unchanged. Record the reason, owner and review date using existing tools. Do not introduce automatic writes to the building system.
At weekly reviews, count exceptions with clear ownership, reviews completed when intended and cases where missing context required another visit or conversation. Record cases where the exception remained necessary. Compare the effort with the usefulness of the information, and retain confusing cases rather than counting only successful closures.
This trial could inform workflow design. It would not prove annual energy savings, establish a predictive model or demonstrate that automated controls are safe. Those questions require different evaluation methods and appropriate engineering oversight.
6. Discussion and research direction
Where does the reason for a temporary building change live today? What evidence should distinguish an authorised exception from a condition that needs investigation?
Logic Foundry is researching sensor anomaly detection, BMS API integration and future reporting and decision tools. One question we want to better understand is how operating intent can accompany building data without overwhelming the people responsible for it. These are research directions, not claims of tested integrations, autonomous controls or measured results.