AI and Power Engineering - Part 6 | AI Interaction Modes

AI interaction Modes

Most discussions of “AI in engineering” collapse the technology into a single thing - usually a chat window. That misses most of what the technology actually does. There are a number of distinct ways of using AI in power-systems work.

1. Question and Response

This is the most basic interface that most users will be familiar with - a simple chat-style window. You ask the AI a question, it responds. That can be a quick one-shot answer or a long iterative session where reference documents, results and model outputs are uploaded and analysed. This is a simple an easy method, but easy to dive into rabbit holes.

2. Inbuilt AI

This is perhaps the easiest mode to understand, but one of the hardest to assess properly. Many software packages now claim to have AI built in. That can mean several different things. In some cases it is a genuinely useful assistant built around the software environment. In other cases it is a wrapper around a chatbot, a documentation search tool, or a conventional optimisation algorithm that has been rebranded. The terminology is loose and the marketing tends to run ahead of the engineering.

For power-systems software, the key question is what the AI can actually see and do. Can it access the active model? Can it inspect parameters? Can it understand study cases? Can it check results? Can it explain which assumptions it has used? Can it be audited?

3. Direct interface

A more powerful approach is a direct interface between the AI and the engineering software. Model Context Protocol (MCP) is one example. In plain terms, MCP lets an AI system connect to external tools and applications in a controlled way. For power systems work, this means an AI interface to software like DIgSILENT PowerFactory, PSCAD or ETAP. The AI does not just talk about the model - it can interrogate it. It can call scripts, inspect objects, list study cases, read parameters, export results, and help automate repetitive workflows. Building and developing an MCP interface is not simple however, it takes time to create and build the toolcalls and skills that are necessary.

4. Agentic

The next level beyond a direct interface is an agentic workflow. The AI is not simply answering questions or calling one tool at a time, but is able to break a larger task into steps and carry them out through a set of tools, scripts or sub-agents. This is where the technology becomes genuinely interesting for power-systems engineering. An AI agent can be asked to inspect a model, identify relevant plant, check controller parameters, export results, compare plots, find inconsistencies, generate test cases, and produce a structured summary of what it found. This is also where things can go horribly wrong.

Summary

A key step in the AI journey with power systems analysis is understanding your use case and where AI can help, where it will not and how you can use it for specific tasks.

I am planning to start writing some deep dive articles and show detailed use cases. I have dropped a small poll below - let me know what you think!