A strategic proof of concept

Become an AI company without owning the LLM.

Navigator demonstrates how specialized risk intelligence can be made available inside the AI environments utilities already choose, while creating a practical path from technology-enabled services to scalable software.

Explore the strategic case Try the connection

The investment thesis

Own the intelligence. Meet customers in the AI they already use.

The large language model market will keep changing. A specialized software company does not need to predict which model wins. A Model Context Protocol (MCP) server is a lightweight program that gives artificial intelligence applications a standard way to connect to external data sources, tools, and services with defined guardrails. It makes specialized expertise usable across the models and operating environments customers select.

01 / ENABLE AI WITHOUT OWNING THE MODEL

Put specialized intelligence inside the utility's AI strategy.

An MCP server gives authorized agents a governed way to discover and use specialized capabilities. It can support ChatGPT, Claude, Copilot, Esri AI assistants, and utility-built agents without creating another general-purpose AI destination customers must adopt. It can also give an existing AI application the context and instructions needed to use MCP servers connected to other enterprise systems, when those connections are authorized, so it can assemble a broader operating picture without building another set of point-to-point integrations.

  • Reduce avoidable cyber exposure. Keep governed tools, permissions, and system boundaries between operational data and the language model.
  • Lower development and lifecycle cost. Invest in the context layer instead of continually rebuilding the conversational interface as models change.
  • Meet utilities where they operate. Make the same trusted intelligence available across engineering, integrity, finance, regulatory, and executive workflows.
  • Preserve customer choice. Let each utility select its preferred model, cloud, security architecture, and user experience.
02 / BUILD THE PATH TO TRUE SAAS

Turn expertise delivered by people into capability delivered by software.

Specialists will remain essential, but dependence on them for repeatable work can be reduced. MCP creates a way to encode established methods, decision context, interpretation, and guardrails into software that utility staff can use every day. The result functions like continual on-the-job training, reinforcing sound practices in the flow of work instead of limiting knowledge transfer to a point-in-time class or consulting engagement.

  • Productize the operating model. Package model outputs, interpretation, workflows, and guardrails as discoverable capabilities.
  • Scale without matching headcount to usage. Serve more users, functions, and decisions from the same governed product layer.
  • Create recurring customer value. Move from periodic analysis to intelligence embedded in planning and operating decisions.
  • Elevate the experts. Shift scarce talent toward model innovation, complex judgment, and customer outcomes rather than repeated data assembly.

The durable advantage

The model is valuable. The context around the model is the moat.

Utilities do not only need a risk score. They need to understand what drives it, what data supports it, how it changes over time, which actions reduce it, and how to explain the decision consistently to engineers, executives, boards, and regulators.

“The opportunity is not to compete with the companies building language models. It is to become the trusted context layer those models call when the decision involves a complex gas system.”

Strategic premise demonstrated by Navigator
Utility AI environmentChatGPT, Claude, Copilot, or an internal agent selected by the customer
MCP context layerGoverned tools, context, definitions, evidence, workflows, and permissions
Models and enterprise systemsRisk models, GIS, SAP, EAM, leak, work, customer, and planning data

Sample prompts and outputs

See how one connection supports three levels of work.

The final prompts and generated documents are coming next. These placeholders show the intended experience and output structure.

Executive decision

Board Deck

Turn portfolio, risk, regulatory, physical-scope, and scenario data into a concise decision package for senior leadership and the board.

View prompt and output
Model evolution

Adding Threats

Define a new risk threat or signal, identify required data and systems, and create a governed path from concept through validation and deployment.

View prompt and output
Engineering workflow

Plan a Project

Move from segment-level risk to a practical project scope, conceptual estimate, utility map, implementation warnings, and simulated work order.

View prompt and output
Start with a broad question

What can I do with this Navigator connection?

Experience the proof of concept

Connect Navigator to the agent you already use.

The same specialized capability layer can be discovered by multiple AI environments. This demonstration is remote and stateless; availability and administrative approval depend on account and workspace settings.

ChatGPT custom app

Add Navigator as a custom MCP app through developer mode.

  1. Open Workspace settings, then Apps.
  2. Enable developer mode if your workspace permits it, then select Create.
  3. Name the app Navigator and enter http://navigator.dataoperationalized.com/mcp as the MCP server URL.
  4. Review the discovered tools, create the app, and enable it in a new chat.