From Requirements to a Hardware Model with Jama Connect MCP
See how governed requirements guide AI-assisted CAD modeling, verification, and design changes.
Turning system requirements into a hardware design often requires hours of manual modeling, analysis, and verification. With Jama Connect and Model Context Protocol (MCP), engineering teams can give AI agents access to governed product requirements to accelerate CAD model creation, evaluate designs against specifications, and maintain traceability as requirements change.
In this demo, Patrick Knowles, Manager of Solutions and Consulting for Aerospace and Defense at Jama Software, demonstrates how Jama Connect, OpenAI Codex, and Autodesk Fusion 360 work together to create and refine a model rocket assembly. See how AI uses requirements managed in Jama Connect to generate CAD models, incorporate commercial off-the-shelf components, perform mass analysis, and produce implementation summaries and verification results.
The demonstrated workflow completes these activities in minutes, with engineers reviewing AI-generated outputs to confirm alignment with design requirements.
Transcript
Connecting requirements to hardware design
Hello, and welcome. My name is Patrick Knowles, and I’m the Manager of Solutions and Consulting for Aerospace and Defense here at Jama Software. Today, I’m going to walk you through how you can go from requirements to a hardware model in just minutes using Jama Connect’s Model Context Protocol integration with Fusion 360 or another CAD tool of your choosing.
The benefits include reducing the time it takes to move from requirements development to initial drafts and designs in CAD, as well as automating verifications throughout the process.
On the screen, you’ll see Jama Connect. There’s a list of requirements on the right, and you can see the final model we’ll look at later in the rich text box on the left.
Defining the AI first task
This is all done through what’s called an AI-first task, or AFT. This execution prompt is my AFT. It outlines the information we’ll gather, what we’ll create, and how we’ll provide the prompt to Fusion 360. Again, this approach can be used with other tools.
It also covers the analysis of the created model, the implementation summary, and the relationships back to the items in Jama Connect.
Reviewing the implementation summary
What exactly is an implementation summary? Here, we have three of them: my initial run and two more where we modified the rocket and then added a commercial off-the-shelf part. Let’s look at the original update to the rocket.
This model rocket has a simple design with standard features you’d expect to see. It also includes an analysis of the space available for a payload.
The summary lists the features and requirements considered in the design and indicates whether each was incorporated. You can see that all of them were used in this run.
You’ll also see a mass budget with estimated masses, mass allowances, and the related requirements. Based on the material selected in the CAD tool, the analysis shows the calculated mass of the different objects.
There’s additional analysis of the features and requirements, which we’ll explore in more depth in the itemized view. For commercial off-the-shelf parts, the summary can even include a direct link to the manufacturer’s technical specification.
Maintaining traceability and human review
Why is this important to do in Jama Connect? You get Live Traceability, which connects the information from the original generation of the solid model to the requirements and verifications.
Here, you can see an automated run completed through my automated test plan, as well as the manual run I performed afterward to confirm that everything was correct and modeled properly.
Review Center in Jama Connect also supports this process, keeping a human in the loop throughout development and enabling review of the models and other outputs against manufacturing standards and requirements.
Generating and inspecting the CAD model
This workflow uses agentic coding. We’re creating a text-based model rather than traditional code, but the approach is similar to spec-driven development.
Here, I asked Codex to execute the prompt we looked at earlier. I also told it to put the information in this folder, which is open on the left.
Along with the information that supports the implementation summary, the primary output is the information Fusion 360 ingests to create the model based on the product requirements and system features.
This leads us to a complete model that we can use, review, inspect from different views, and investigate. In this view, we see a smooth-bodied object, which is one of the requirements, along with the colors specified in the design.
If we go back to the wireframe view, we can see the internal motor mounting tube and the C6-5 motor, which we identified as a commercial off-the-shelf, or COTS, part required for this design.
Updating the assembly with a COTS component
Earlier, I mentioned that I needed to update this model to include a new part, so I asked it to generate an assembly. Here is my rocket, and here is my assembly. The assembly includes a commercial off-the-shelf rocket stand for launching the system.
During this process, the analysis showed that my original design did not account for the difference in diameter between the payload section and the primary structure. We generated an additional piece so the rocket could launch from this stand without interference from the payload structure.
That’s how we arrived at this assembly. The summary also identifies the feature and requirement I hadn’t included previously.
These new requirements and features were added and used to update the entire model, along with the related analysis.
From requirements to a model in under 20 minutes
In conclusion, we can go from requirements to a hardware model in just minutes. The run I performed took under 20 minutes to generate all the hardware models we see here in Fusion 360 and return the verification results.
Afterward, a human could test and finalize the system as a second check on the generated results. That keeps human governance in place to help ensure our products are good and safe.
This workflow is enabled by Jama Connect’s Model Context Protocol integration and connections to tools such as Fusion 360 through MCP.
