Agentic grid planning on AWS
Program built in collaboration with Duke Energy and managed by AWS. Your engineers stay in control while AI agents compress months of study work into minutes.
Meet growing demand for interconnection studies
Utilities are being asked to run more studies, faster, with more scenarios and deeper analysis, but engineering capacity has not kept pace. A large share of each study (preparing data, moving it between tools, updating cases, rerunning procedures, and assembling reports) needs engineering oversight, not continuous engineering judgment.
Prepare cases, run analyses, and rerun studies with AI agents
Agentic Grid Planning on AWS gives utilities access to AI agents, developed with Duke Energy and managed by AWS, that run study workflows using the physics-based simulation software, grid models, scripts, and standards their teams already trust. This enables the engineers to complete studies faster, analyze more scenarios, and focus on judgment and decisions, with a versioned, reviewable record throughout. The agents support steady-state transmission studies for generation and large-load interconnection using power flow and contingency analysis, for a single request or a cluster.
Complete studies faster
Help your team finish studies faster, explore more scenarios, and spend engineering time on judgment and decisions rather than repetitive setup.
Use the tools you trust
The AI agents work with your existing simulation software, models, and scripts. The software does the deterministic modeling; the agents run and coordinate the workflow.
Keep a reviewable record
Agents version artifacts and record every step, so reviewers can separate simulation evidence, AI reasoning, and engineering decisions, and trace values to their runs.
What you can do
Agentic grid planning on AWS automates the three most time-consuming stages of a transmission interconnection study, with engineer review checkpoints built in at every decision gate.
Prepare study-ready cases
Select base cases and a single interconnection request or a cluster of requests. For each request, the AI agents propose a point of interconnection and the topology updates needed to connect it. Your engineers confirm or override the recommendations before the agents generate updated cases across planning scenarios.
Run and rerun steady-state studies
Run power flow and contingency analysis, categorize violations, propose mitigations for approval, and draft reports. When inputs change, rerun without rebuilding the study by hand.
Capture and reuse procedures
Direct the first study interactively, then have the AI agents draft a workflow guide from the recorded session. Your engineers review and promote the guide for reuse, turning experienced engineers' methods into a versioned team asset.
Secure by design — your data, your control
Agentic grid planning on AWS was co-developed with leading transmission utilities and built to the AWS standard for security. Your data is controlled by you, using the same AWS mechanisms your teams already run.
Your environment is attached to your own AWS account. All resources run in your infrastructure, not shared.
Access is governed by IAM policies and your own identity provider through AWS IAM Identity Center. You define who can do what.
The agents' reasoning runs on foundation models through Amazon Bedrock. Your prompts and completions are not used to train models and are not shared with model providers.
FedRAMP-aligned, SOC 2 Type II, ISO 27001 certified. Built for the regulatory requirements of energy and utilities.
Learn how utilities are accelerating interconnection studies
Interested in what agentic grid planning could do for your queue? Connect with our Energy & Utilities team to explore how this co-developed solution can work for your study process.
