{"resourceId":"smud-sd10-innovation-monitoring-ai-meters-robots-2026","versions":[{"version":"external-a106a41c63c67636c4ba5aae3dc573426758512db41e0ffe7747489cf14e38fd","resource":{"id":"smud-sd10-innovation-monitoring-ai-meters-robots-2026","title":"SMUD SD-10 Innovation monitoring report: AI fire detection, robot substation inspection, 100+ custom GPT models and next-generation meters","organization":"Sacramento Municipal Utility District (SMUD)","sector":"Municipal public power utility","geography":"Sacramento, California, United States","publishedAt":"Exhibit to Agenda Item 2, Board Policy Committee and Special SMUD Board meeting, 7 Oct 2026","publicationDate":"2026-10-07","eventDate":"2026-10-07","sourceName":"SMUD Board of Directors","sourceLabel":"Public utility board monitoring report exhibit (PDF)","sourceUrl":"https://www.smud.org/-/media/Documents/Corporate/About-Us/Board-Meetings-and-Agendas/2026/Oct/Exhibit-to-Agenda-Item-2--SD10-Innovation.ashx","evidenceClass":"government-evaluation","outcomeClass":"emerging","topics":["infrastructure","knowledge-work","governance-procurement"],"finding":"SMUD's annual board monitoring report on its Innovation strategic direction lists 28 R&D projects in the 2026 report (up from 22 and 25). It describes wide-area AI-enabled fire detection (NextPower Vantage at the Rancho Seco generation site, using AI-based visual processing to spot nearby fires), ANYbotics ANYmal robots paired with substation inspectors for infrared imaging (57 of 270 substations commissioned), staff-built 'SMUD GPT' custom models (over 100, described as increasing internal efficiencies), a new outage management module integrated with existing ADMS/DERMS, and next-generation meters described as the foundation for distributed intelligence applications. Management asked the board to accept the report.","sledRelevance":"A public power board's own record of how AI and automation are being governed: reported to elected directors against a written strategic direction. It covers meters, asset inspection, wildfire risk and internal generative AI in one place.","evidence":"Official board exhibit (slides). Claims are management-reported counts and descriptions; no accuracy, cost or reliability outcomes are given for AI tools.","architectureImplications":"Next-generation meters treated as an edge-computing platform; OMS integrated with ADMS/DERMS; AI vision at a generation site and robots for substation imaging.","governanceImplications":"Board-level monitoring against Strategic Direction SD-10 provides a recurring accountability point for innovation spending and AI pilots.","securityPrivacyImplications":"Grid-edge meter intelligence and robots in substations widen the OT attack surface; a future robot pilot for substation security inspection is listed. Controls are not described.","caveats":"Slide format with little detail; '100 custom models' is unexplained and unmeasured. Dated 7 Oct, before this stream's first edition, and not archived; included now because it is a primary public-power record covering metering and AI, an area thin in both earlier editions.","streamIds":["utility-management"],"roles":{"sales":"Sales framing: SMUD reports AI and robotics to its elected board under a written innovation policy. That gives a template for public power utilities that need board-level accountability for AI pilots. Open with how AI work is reported to the board today and whether outcome measures exist. A bounded engagement is a board-ready AI portfolio report with measures for each pilot. Do not claim efficiency gains from SMUD's custom models or reliability benefits from robots, as none are quantified.","engineering":"Engineering framing: several AI threads share data dependencies. Next-generation meters as a distributed intelligence platform, an OMS integrated with ADMS/DERMS, AI vision for fire detection and robot infrared imaging all produce operational data that needs to be joined with GIS and asset records. A proof of value could measure robot inspection findings against manual inspections, or fire detection alerts against confirmed events. Internal GPT models need data access controls.","delivery":"Delivery framing: the owner is the innovation or R&D lead reporting to the board, with grid operations and IT security. Work includes defining outcome measures for each pilot and an inventory of the custom generative AI models staff have built. Acceptance should be proposed per-pilot measures in the next monitoring report. Risks include unmanaged growth in staff-built models, OT exposure from new edge devices, and reporting activity counts instead of outcomes."},"roleDetails":{"sales":{"customerProblem":"Public power boards approve innovation spending but receive activity counts rather than measured outcomes for AI pilots.","stakeholders":["Board of directors","General manager","Chief innovation or R&D officer","Grid operations","Chief information security officer"],"discoveryQuestions":["How are AI pilots reported to the board today?","Is there an inventory of staff-built generative AI models?","What outcome measures exist for robot or AI inspection pilots?"],"valueHypothesis":"Outcome-based reporting could help boards decide which AI pilots to scale.","potentialEngagement":"A board-ready AI portfolio review with an inventory and outcome measures for each pilot.","unsupportedClaims":["Efficiency gains from custom models","Reliability improvements from robots","Fire detection accuracy","Savings from next-generation meters"]},"engineering":{"fit":"Fits large public power utilities with OMS/ADMS/DERMS and AMI modernization underway.","architectureIntegration":"Edge-intelligent meters, OMS integrated with ADMS/DERMS, AI vision at generation sites, robotic substation imaging, internal generative AI models.","prerequisites":["AMI 2.0 meters","GIS accuracy","Asset records linked to inspection data"],"constraints":["Little published performance data","OT security for new edge devices","Model sprawl"],"security":"Segment robots and edge meters from critical control networks and govern data access for internal GPT models.","proofOfValue":["Compare robot inspection findings with manual inspections","Track fire detection alerts versus confirmed events","Measure time saved by top custom models"]},"delivery":{"work":"Pilot measure definitions, generative AI model inventory, security review of edge devices and robots.","dependencies":["R&D team","Grid operations","IT and OT security","Board reporting cycle"],"ownership":"Innovation or R&D lead reporting to the board.","skillsAdoption":"Substation inspectors work alongside robots; staff building custom models need guidance and review.","governanceCheckpoints":["Annual SD-10 monitoring report","Security review before new edge pilots","Model inventory review"],"acceptanceCriteria":["Proposed: each AI pilot reports an outcome measure in the next monitoring report","Proposed: complete inventory of custom models with owners","Proposed: security review completed for robots and edge meters"],"risks":["Model sprawl","OT exposure","Activity reporting without outcomes"]}},"retrievedAt":"2026-10-31T22:07:01-05:00","enrichedAt":"2026-10-31T22:14:58-05:00","enrichmentBasis":"retrieved source","accessibilityWorkforceImplications":"Robots are paired with inspectors rather than replacing them; staff are building their own AI models, implying a need for training and support.","procurementImplications":"Mix of vendor products (NextPower, ANYbotics) and in-house models; EPRI and grant-funded work share R&D cost.","operatingModelImplications":"Innovation is governed through a board strategic direction with annual monitoring.","sourceVerification":{"openedUrl":"https://www.smud.org/-/media/Documents/Corporate/About-Us/Board-Meetings-and-Agendas/2026/Oct/Exhibit-to-Agenda-Item-2--SD10-Innovation.ashx","referenceExcerpt":"Staff have developed over 100 custom models","promptVersion":"sled-research-v4.0","model":null,"basis":"agent-reported inspection"}}}]}