AI Implementation for Energy & Utilities in Plano, TX

Oncor Electric Delivery serves Plano as part of its 13-million-meter North Texas footprint — the largest T&D utility in Texas by meter count, owned by Sempra and Texas PUF investors, regulated by the Public Utility Commission of Texas, operating inside ERCOT. The Plano service territory carries a customer-mix distortion that matters for any AI work: a heavy concentration of Class 4 and 5 commercial accounts relative to residential, which drives load factors, demand-response participation rates, and outage-response SLA expectations closer to what you'd see in a downtown core than a typical suburb.

Legacy West alone changed the feeder topology conversation. The campus buildout pulled feeders into a network configuration more typical of dense urban load, and the ongoing Toyota, FedEx Office, Liberty Mutual expansions keep pushing transformer loading into near-capacity regimes that AI-assisted load forecasting and transformer health analytics can genuinely improve. North Texas data center load growth — the larger regional story driven by hyperscale buildout west of DFW — doesn't hit Plano directly the way it hits Irving or Richardson, but it raises the stakes on every capacity planning decision Oncor makes inside the region.

Weather exposure is the other operational variable that shapes AI scoping. Uri-scale freeze events are now a permanent part of ERCOT planning reality, and Plano's 2021 experience with load-shed rotations under the EEA-3 order is institutional memory for every corporate-campus facilities lead in the city. May-September convective season drives the SAIDI/SAIFI numbers that matter in PUCT reporting. AI systems that don't load-test against Uri-class cold snaps and Derecho-class storm cells don't survive a North Texas production cycle.

MSG is 289 miles southeast of Plano on I-45 and US-75 — roughly a four-and-a-half-hour drive. That's a multi-day-visit commitment rather than a day-trip, which we scope into engagement structure: 3-4 day kickoff immersion, onsite cadence pegged to integration milestones, and pre-summer-peak onsite readiness review in late May.

Against Oncor's operational stack, the highest-leverage first AI systems for a Plano-focused engagement cluster around three patterns. First, OMS triage tuned for the coincidence of Legacy West campus outage calls — a single tenant losing power generates 400+ parallel customer-service contacts within the first 15 minutes, and the triage system needs to collapse those into a correct feeder-level incident without dropping a legitimate second incident on a different feeder. Second, AMI analytics that actually leave MDMS — Oncor's Itron/Landis+Gyr deployment produces meter-level voltage, outage-flag, and interval-load data that in most utilities still lives in a Meter Data Management System read by a billing analyst once a month. We build retrieval and inference layers that pull that data into operational signal: transformer-loading anomaly detection, non-technical loss pattern identification at the commercial meter level, voltage excursion alerts at the service drop. Third, demand-response and 4CP forecast accuracy improvement for the commercial programs Oncor and the competitive retail providers operate against — Plano's Class 4/5 participation in those programs is heavy, and the economic value of tightening coincident-peak forecasts inside a 30-minute window is measurable.

Integration work is non-negotiable. Oncor operates an industry-standard ADMS stack with GE/Schneider patterns, Esri ArcGIS Utility Network for GIS, Oracle CC&B on the CIS side, and a mature AMI headend. Every AI system we build reads through governed data contracts — AF extracts, ODS pulls, API layers IT owns — never directly into BES Cyber Assets or control-plane systems. Retrieval and inference run inside your VPC and CIP perimeter where data classification requires it. Evaluation harnesses use Oncor's real historical operational data including Uri-week and 2023 Derecho event data — not synthetic benchmarks. Deterministic fallbacks are mandatory on anything that touches operational decision support. Handoff documentation is structured for Oncor's IT, ops, and reg-affairs teams to own and audit at month 18.

Texas utility AI has a specific regulatory shape that shouldn't be glossed over. ERCOT operates outside FERC jurisdiction for most purposes, which changes the reliability-standards conversation — NERC CIP still applies to BES Cyber Assets, but the market-layer data exchange with ERCOT has its own protocols and timing requirements that AI systems touching bid submission, ancillary services, or real-time market data need to respect. PUCT oversight covers rate cases, reliability reporting, and prudence review of capital investments. An AI system classified as part of capital-investment spending has to survive PUCT prudence review — the documentation of value has to be structured the way rate-case filings are structured.

Post-Uri reliability is the dominant Texas regulatory conversation of this decade. Every utility capital plan, including AI-adjacent spending, is evaluated against contribution to reliability under extreme weather. Oncor's Uri experience was less catastrophic than some ERCOT entities but the system-wide political and regulatory aftermath — SB 3, weatherization requirements, PUCT market-design reform — reshaped the environment. AI investments that improve storm-event operational performance, load-shed coordination accuracy, or restoration sequencing have a clear path through prudence review. AI investments pitched as abstract modernization don't.

The corporate-campus concentration in Plano adds a customer-experience layer that doesn't exist in most utility markets. Toyota's facilities team, JPMorgan's operations leadership, Liberty Mutual's business continuity group each maintain expectations about outage communication accuracy, ETR precision, and post-event analytics that match what they'd expect from a commercial SaaS vendor. Oncor's customer-facing tools for these accounts — and the AI systems that feed them — operate at that SLA standard, not at a residential-customer-service standard. We scope accordingly.

MSG is a Gulf Coast operator-consulting firm that ships production software and has for the last decade. ServiceStorm operates at multi-tenant SaaS scale across Gulf Coast home services operators through hurricane seasons and freeze events. MFGBase is a B2B marketplace connecting manufacturers. LocalAISource is an AI professionals directory. That's operator experience, not a consulting deck. When we come into an Oncor-adjacent Plano engagement, we bring engineers who know what production reliability means in Texas weather — because our own systems survive Texas weather.

The 289-mile distance from Beaumont is honest. We're not down-the-street the way we are with Houston engagements. But the corporate-campus dynamic in Plano favors an onsite cadence pegged to integration milestones rather than weekly standups — a pattern we've refined across North Texas engagements. Multi-day kickoff immersion, integration-sprint anchoring visits, pre-summer-peak readiness review in May, post-winter-peak lessons-learned review in February.

We refuse scopes that don't ship. Most AI consulting in utility space ends at a roadmap deck. Ours end at a system running at month 18 without us, documented for PUCT prudence review and NERC CIP audit. The trade-off versus a national consulting firm is specific: you get engineers instead of analysts, Gulf Coast operational instinct instead of coastal-city abstractions, and production artifacts instead of slideware. What you don't get is a 50-person onsite team; we scope tight and ship.