AI tooling

Summary

I led development and operationalization of an AI-powered content design model to create content at scale.

Situation

Content designers were stretched, leading to developers and designers creating their own, inconsistent content.

Problem for Autodesk

Content quality and style varied across products, making the experience feel uneven and hard to scale.

Problem for content creators

People lacked a simple, reliable way to create content that followed our guidelines, so work took longer and results were inconsistent.

Approach

  1. Ran a proof of concept pilot with a standalone team

  2. Defined a simple, scalable architecture

  3. Prioritised features and improvements based on real team needs

  4. Ran user research to understand needs and build scalable onboarding approach

Solution and rationale

  • Dual model approach with a Cursor-based workflow and a Toolkit app

  • Integration with MCP-servers, shared rules, and examples

  • AI outputs grounded in real Autodesk context

  • Scheduled testing and reporting to maintain quality

Impact

  • Thirty teams using the Toolkit to self-support on content creation

  • Tasks now take minutes instead of hours

  • Automation of entire workflows

  • Content quality improved