Using the Rescale Assistant

Overview

The Rescale Assistant is an AI-powered chat interface built into the Rescale platform. You can ask it questions in plain language to query your job data, summarize results, search across your connected data sources, and take common platform actions without leaving your current screen.

The Assistant is available from anywhere in the platform and understands the context of the job and simulation life-cycle, you can ask follow-up questions the way you would with a colleague.

Highlights

  • Natural-language access to your data — Ask about job status, results, errors, and history without writing a query or clicking through menus.
  • Job summaries on demand — Get a plain-language summary of what a job did, including errors surfaced from the process output, so you don’t have to read raw logs.
  • Streaming responses with “Show Work” — Answers stream in as they’re generated, and you can expand the Assistant’s reasoning to understand *why* it gave the answer it did.
  • Access specialized agents — Reference pre-built agents to complete more complex tasks like input deck validation, job troubleshooting, and report generation.

Using the Rescale Assistant

This section describes how to open the Assistant, ask questions, and use its core capabilities.

Prerequisites

Before you begin:

  • Your organization must have Advanced Modeling and Simulation or Agentic Engineering features enabled. Access is controlled by your workspace administrator.
  • You need an active Rescale account with access to at least one workspace.
  • To ask about a specific job, you need permission to view that job.

Open the Assistant

  1. Click the Assistant icon in the platform navigation to open the Assistant drawer.
  2. The drawer opens alongside your current page. You can keep working in the platform while the Assistant is open.
  3. To switch to a larger view, click the full-screen control at the top of the drawer.

Tip: Anything you’ve typed but not yet sent is preserved if you close and reopen the drawer, so you won’t lose a question in progress.

Get Started Tips

If you’re not sure what to ask, click the ? icon at the top of the Assistant. This sends the question “What can the assistant do?” on your behalf, and the Assistant responds with an overview of the questions and tasks it can help with.

Clear the Conversation

To start fresh, click the Clear button at the top of the Assistant. This permanently deletes the current conversation and starts a new one.

Note: Clearing a conversation cannot be undone. The messages are removed permanently and can’t be recovered.

Ask a Question

  1. Type your question into the message field at the bottom of the Assistant.
  2. Press Enter or click Send.
  3. The response streams in as it’s generated. You can start reading before the full answer is complete.

Example questions you can ask:

  • “What’s the status of my most recent job?”
  • “Why did job *<job name>* fail?”
  • “Summarize the results of this job.”
  • “Show me jobs that ran ANSYS Fluent in the last week.”

Summarize a Job

The Assistant can read a job’s output and give you a plain-language summary, including any errors found in the process output file.

  1. Open the job you want to summarize, or name it in your question.
  2. Ask the Assistant to summarize it (for example, “Summarize this job”).
  3. Review the summary. If the job encountered errors, the Assistant surfaces them so you can act without opening the raw logs.

See the Assistant’s Reasoning (“Show Work”)

When the Assistant works through a multi-step question, it shows the tools it’s using while it works. Once a step finishes, that step is cleared from view to keep the response readable.

  1. Send a question that requires the Assistant to look something up or take several steps.
  2. Watch the in-progress steps appear as the Assistant works.
  3. To understand *why* you got a particular answer, expand the Show Work view for a narrative explanation of the Assistant’s reasoning. This focuses on the reasoning, not low-level technical detail.

Use a Specialized Agent

For tasks that go beyond a single question, the Assistant can hand off to a specialized agent. Availability depends on your workspace configuration.

  • Input deck validator — Checks your solver input settings before you run, catching configuration problems early.
  • Job troubleshooter — Helps diagnose why a job failed and suggests next steps.
  • Report generator — Produces a formatted report from your job results.

To use an agent, describe what you want in natural language (for example, “Validate my input deck” or “Help me troubleshoot this failed job”), and the Assistant routes your request to the appropriate agent.

Note: If an agent isn’t available in your workspace, contact your workspace administrator about enabling advanced AI features.

FAQ

Do I lose my question if I close the Assistant? No. Any text you’ve typed but not yet sent is preserved when you close and reopen the Assistant drawer.

Can the Assistant see jobs I don’t have access to? No. The Assistant respects your existing permissions. It can only access jobs and data you’re already allowed to view.

Why does the Assistant sometimes show steps that then disappear? The Assistant shows each tool while it’s in use so you can see progress. Once a step completes, it’s removed to keep the response clean. To review the full reasoning, expand Show Work.

What’s the difference between the Assistant and an agent? The Assistant is the general chat interface for questions and platform actions. Agents are specialized tools (such as the input deck validator or job troubleshooter) that the Assistant can route to for specific, more involved tasks.

Next Steps

Relevant Links

×