The Error Project

We: Licensing team struggled to define which of thousands of errors to improve first. Engineering leads insisted that the PM look into which error events were occurring most frequently within Autodesk's biggest products.

+Me: Noticing the most frequently occurring errors did not overlap with those responsible for 20% of customer support cases, I proposed the first errors we address be a balance of technical and user problems, leading the team to focus on 2 top-occurring errors and 3 of Support's top errors.

We: The PM sought buy-in from product teams to accept code changes for a new framework, so that the Licensing team could modify licensing-related error experiences. Developers dug into legacy code in search of root causes. The technical solution lead explored how to feed new error messages to products with and without a connection to the Autodesk Desktop Licensing Service.

+Me: I explored the elements of a good error experience, the variety of errors across Autodesk, existing templates and UI styles. I also worked with the technical solution lead to understand what can be passed and surfaced.

We: The PM, technical solution lead, and UX (me) determined what an ideal error experience looks like when all conditions are met, and the fallback experience for when the Autodesk Desktop Licensing Service is unreachable.

+Me: I created a template for errors that could deliver the same key elements. I sourced visual components from the Autodesk design system and reviewed designs with the Visual Design team, while using Microsoft's safe dialog components for the fallback experience. I co-created copy-text with a senior content designer to ensure consistency with Autodesk language and tone.

We: The PM and Development leadership wanted a guarantee that the new error messages would indeed be "improvements" before committing resources. While I user-tested new error experiences, developers began building the UI template with placeholder text and testing the solution's efficiency. The PM chose checkpoints each product team's willingness to adopt the framework.

+Me: I tested the original errors as a baseline with un-moderated think-aloud tests, against which I tested the new errors with quantitative and qualitative insights. I identified user misinterpretations, expectations, and reasons for low ratings; this helped me to rewrite error messages and share data-based do's and don'ts with the Content Design team.

We: Developers replaced placeholder text within final error messaging, and PM and Engineering leads worked with Product leads to absorb the new framework.

+Me: I held several review sessions with Visual Design, Content Design, and UX designers from Product teams to consider their feedback and seek cross-team approval.

We: Whenever a product team absorbed the new error framework, the PM monitored data, such as user activity on the new Licensing-owned error dialogs, technical error events, and customer support cases. In response to product teams' requests, the PM and solution lead also discussed adapting the framework for macOS.

+Me: I requested that user interactions with new licensing-owned error dialogs be visualised on a dashboard to understand the volume of customers who effectively click on the troubleshooting hyperlink versus those who click on other elements on the dialog. I also proposed how the UI template might translate into a macOS dialog in the future, and when the Visual Design team announced a new design system in 2025, I created a mock-up for how Licensing errors might look when Licensing team adopts the new design tokens.