{"resourceId":"inpt-energy-policy-forecasting-savings-distinction-2026","versions":[{"version":"external-d4c7de2a14af9100d84d954ab8aa9dab90a937ac2018989c18d925bafbfccc74","resource":{"id":"inpt-energy-policy-forecasting-savings-distinction-2026","title":"Campus energy brief separates forecasting capability from projected operating savings","organization":"Policy Center for the New South; Imad Hajjaji","sector":"Higher education facilities","geography":"Morocco; conditional transfer to U.S. campuses","publishedAt":"July 8, 2026","publicationDate":"2026-07-08","eventDate":null,"sourceName":"Policy Center for the New South","sourceLabel":"Author policy brief drawing on research; underlying forecasting paper not fully accessible","sourceUrl":"https://www.policycenter.ma/publications/reactive-predictive-how-ai-driven-energy-management-can-transform-university-campuses","evidenceClass":"standards-guidance","outcomeClass":"emerging","topics":["infrastructure","governance-procurement","operating-model","accessibility-workforce"],"finding":"The brief advocates predictive campus energy management, but its energy and financial savings are estimates rather than measured intervention outcomes.","sledRelevance":"Interpretation: a facilities planning comparator for U.S. campuses, subject to different tariffs, building systems, climate and staffing.","evidence":"The author describes INPT metering and forecasting work and explicitly labels the savings section Estimated Impact. It applies a benchmark-based 5–10% savings assumption. This is not an observed reduction attributable to AI. The brief identifies expertise, metering and ownership barriers.","architectureImplications":"Interpretation: prototype a telemetry-to-forecast pipeline in shadow mode before allowing control-system writes; generative copilots are not the technique evaluated.","governanceImplications":"Interpretation: require separate approvals for forecasting, operator advice and automatic actuation.","securityPrivacyImplications":"Interpretation: isolate operational technology, authenticate telemetry and restrict access to occupancy-sensitive traces.","caveats":"Normative brief, not an independent replication. Underlying publisher paper returned 403; accuracy metrics, split methodology and baseline tables were not verified and are not adopted here. No causal savings, total lifecycle cost or U.S. transfer effect established.","streamIds":["campus-operations"],"roles":{"sales":"Interpretation: ask the facilities director, energy manager, finance team and campus IT which demand problem is expensive enough to investigate. What metering exists, and which comfort or laboratory constraints cannot change? A bounded engagement could assess one building's data readiness and define a shadow forecast trial. The value hypothesis is better planning before equipment control is automated. Treat the brief's estimated savings as a question for local measurement. Do not promise transferable percentages, a short payback period or reduced staffing.","engineering":"Interpretation: require trustworthy timestamps, meter calibration records, occupancy calendars and an approved baseline before choosing a model. Compare a simple schedule-based forecast with the candidate using a later held-out period that includes closures and term changes. Test missing-data handling and isolate the experiment from building controls. Proposed validation should report forecast error by operating regime and then separately test whether an operator action changes consumption without degrading service. Resolve the inaccessible paper's methods before using its performance as a benchmark.","delivery":"Interpretation: assign the energy manager operational ownership, with controls technicians and data engineers responsible for integration. Establish a commissioning checklist, incident escalation and a staffed maintenance plan. Train operators using accessible displays and rehearse loss of telemetry. Proposed acceptance requires complete provenance for test data, a documented baseline comparison and a working manual fallback; any later savings claim must use metered intervention results and include total effort. Main risks are seasonal bias, unsafe control changes and an unsupported prototype becoming a critical service."},"retrievedAt":"2026-09-09T03:01:11Z","enrichedAt":"2026-09-09T03:04:27Z","enrichmentBasis":"retrieved source","accessibilityWorkforceImplications":"Interpretation: budget operator training and accessible alarms; do not assume an inexpensive prototype eliminates maintenance work.","procurementImplications":"Interpretation: compare installation, calibration, support and replacement costs, not hardware prices alone.","operatingModelImplications":"Interpretation: name a facilities service owner and maintain a manual fallback.","updateExplanation":"New URL across all 119 archive records. Older evidence added to address the previous edition's facilities gap; no overnight change or new deployment is claimed.","sourceVerification":{"openedUrl":"https://www.policycenter.ma/publications/reactive-predictive-how-ai-driven-energy-management-can-transform-university-campuses","referenceExcerpt":"High-resolution energy data is a prerequisite for regime-aware forecasting","promptVersion":"sled-research-v3.1","model":null,"basis":"agent-reported inspection"}}}]}