by CAPRA J'NEVA
AI/ML - Autodesk - MCP Servers
Spring 2025 - MCP servers were taking the industry by storm, and Autodesk saw a potential disruption to everything they had built with their API platform. A team had been working on an assistant, but all it could do was answer questions based on documentation - not take action. In the platform team, we could see that customers were already building MCP servers to directly call their APIs, building much more powerful capabilities for AI.
We also knew that if we met this moment and designed our own MCP servers, we could much better access the functionality our customers want from agents in their fields. A good MCP server is built on a strong use case, a deep understanding of how customers talk about their problems and what they would prompt.
After managing for 9 years, I was ready to move back into an individual contributor role, and I jumped at the opportunity to partner with the company's best architects and strategists to roll out this program and get it to market.
The Team
5 Division effort, 10 working groups, I was sole dedicated designer, partnered with design teams on each product
Timeframe
6 months
My Role
Experience Architect
Impact
6 months
Time from inception to market
4
Number of public MCP Servers
27
New design patterns built for assistant in 3 months
Experiment! Fast! Now deliver!
As the platform architecture team, we constantly look for trends - we started warning our leaders that our AI position was going to seem weak if we did not respond to what we were seeing in the marketplace. They took us seriously, and MCP server experiments ignited across the company like wildfire.
My job was to encourage this experimentation, while also partnering with a team of 10 developer architects from across the company to create the infrastructure we would need to bring solid MCP servers to market by January. It was the end of June.
I partnered with our VP of engineering for manufacturing products to identify projects to support the engineering effort and helped her lead bi-weekly meetings. I wrote wiki articles on best practices.
End to end design
To release the ecosystem of MCP servers we would need to think through every aspect of how they would be experienced, and partner with dozens of teams to pull off the designs.
There were customer considerations:
How would people discover them both in product and in other channels?
How would they be installed?
How would adminstrator controls work?
How would their verbosity be handled in our assistant interfaces?
How would we support our users in reviewing massive amounts of design changes?
How would people experience answers when they crossed our traditional product boundaries?
What would the future look like in 3 to 5 years?
There were delivery considerations:
How could we coordinate internally so that teams understood who should build what capability?
What were our measures of success and minimum requirements for release?
I successfully coordinated among teams so we could announce and do live demos onstage at our large annual customer conference in mid-September. This included partnering intensively with Autodesk's Assistant team, who builds the primary agent interface, and working closely with our design system team to audit patterns and craft solutions for the numerous changes required to support MCP servers.
What did the future look like?
MCP servers and generative AI will enable a whole new way of working with CAD. Autodesk's interfaces are notoriously difficult to learn. My VP challenged me to prototype what the future would look like 3-5 years out. I worked with experts in each industry team to identify meaningful workflows to test. During this time I was also appointed to advise our research team, who were building for 10 years out on a project called JARVIS. I coordinated with them to ensure my research would feed valuable early customer data into their project.
I built and tested a prototype showing 5 different experiences:
Reviews of massive amounts of agent actions (with Michelangelo Caparo)
A 3D canvas with spatial AI context building
Minimal UI that centers the canvas experience
Collaborative, multi-agent experiences
We recognized that working with AI is much like working with human collaborators. At some point soon, we may work with multiple agents as our team.
The reaction to this prototype was strong - people were ready to bring agents into their workflows, and they particularly wanted ways to intuitively build their project context. The canvas could become the way to do so. Another team is now building that product.
Building this prototype led me to take my next venture in the company with the Design & Make Live incubation team.
