Golden Grid AI Playbook

Golden Grid (Ground-truth Ontology Logging Dynamic Events, Nurturing Grid-wise Response, Improvement, and Decisions) is our open-source ontology for distributed energy resources (DERs). Project details: overview, ontology playbook, and the governance charter. Sponsored by Agile Electrification (UCSD Design Lab).
Our team of designers, developers, and researchers wrote this Golden Grid AI Playbook as an interactive way to understand the Golden Grid ontology and how it is used with LLMs. In our first three months, we focus on home energy data, but we will expand to include grid-operator simulations, policy analysis, and more.
This open source work is grounded in the ontology, data, and workflows shared between Balto Energy and Northern Pacific Power to model, install, and finance DERs.

The Challenge: A Missing Translation Layer
The transition to a decentralized, electrified grid is stalled by a simple, profound problem: we don't speak the same language. Utilities, DER providers, regulators, and homeowners all use different terms for the same concepts. This isn't just semantics; it's a structural barrier to progress.

• Data Chaos: We rely on a messy patchwork of data sources—for example, utilities' APIs (JSON), Green Button (XML), and proprietary DER APIs—each with its own format and quirks. Without a robust cleaning and transformation pipeline, this data is nearly unusable.
• Opaque Policy: The rules governing the grid are buried in thousands of dockets and regulatory filings (often PDFs that are hundreds of pages long). Understanding the real-world impact of an incentive structure or tariff change is a herculean task, making it impossible to design policies that effectively encourage cooperation over grid defection.
• Definitional Mismatches: In our world, it means we can't accurately model or value grid contributions when concepts like site load and gross load are used inconsistently or have inconsistent units to perform calculations.
• Fragmented Workflows: utilities, DER vendors, and regulators publish incompatible schemas and tariff language, preventing shared analysis and workflows with the same data.

Why This Exists
We are not trying to publish yet another comprehensive standard. Our aim is to be maximally useful to a small set of influential practitioners who need a shared mental model that travels between datasets and policies. What we accomplish in this initial open source contribution: Build a thin, working ontology. We share examples of how users of this ontology can map their own data (nouns), transforms (actions), and workflows (workflows) with this ontology.

• Composable ontology: We share an open source ontology so collaborators can compare households, policies, and assets without rebuilding pipelines while keeping the data provenance intact. No need to model the world to use this codebase—for example, you can use just the workflows or adapters in your own data pipelines.
• Extendable adapters: We share open APIs and formats so DER vendors, utilities, and regulators can use data and workflows with the ontology. Contribute back through transparent governance. No software lock-in—for example, you can extend or simply use Golden Grid adapters.

Our Philosophy: An Ontology for Communication
Inspired by a key insight from our project kickoff, this ontology is founded on a simple principle: its primary goal is effective communication. Many standards fail because they are rigid and exclusionary. Our approach is different:

1. A Language, Not a Law: The ontology is a vocabulary and a grammar for describing the relationships between energy assets, financial instruments, and regulatory rules. It is designed to be expressive and precise.
2. Open, Adaptable Governance: The ontology is not static. It is designed to evolve through open contribution. When a new technology, metric, or policy emerges, our governance model provides a clear mechanism to extend the language, ensuring it remains relevant and useful. We will accept pull requests to adapt the specification.
3. Bias Through Transparency: Any ontology has values embedded within it. By making the governance process transparent and community-driven, we commit to the value of effective communication, allowing us to identify and integrate new concepts or resolve conflicts, rather than letting bias creep in through obscurity.

Orientation

How might we use the Golden Grid AI Playbook?

Below, we show how layers assemble data, policies, and workflows into question-ready ontologies. Each tab highlights user questions, connective tissue, and the signals that make the support loop work today via CLI tooling—while pointing toward future AI chat or web interfaces that can sit atop the same ontology.