AI-Enabled Skill Validation
Skillable
Product Design Lead: Kim Oswalt
Engineering Lead: James Grau
Product Managers: Eric Grau, Sarah Mullens
Background
Skillable labs let users learn, practice, and validate skills in a virtual environment. To track performance, lab authors use activities—typically scripts that check if a user completed a task correctly.
Time
Writing scripts can be a tedious and time-consuming process for lab authors.
Expertise
Scripting is a specialized skill that requires coding knowledge and domain expertise.
Limitations
Certain tasks, like confirming a toggle is switched on, can't be reliably handled with scripts.
We saw an opportunity to make lab building faster and more accessible by harnessing AI's ability to interpret what's on a user's screen.
AI Vision Activities let authors write plain-language prompts instead of scripts. Behind the scenes, AI captures a snapshot of the user's environment and checks it against the expected outcome—no code required.
Design Approach
I partnered with engineering and product to craft a simple but powerful workflow that:
01
Supports new & experienced lab authors
02
Offers just-in-time guidance for writing effective prompts
03
Prioritizes clarity, confidence, and iteration
Integrated test functionality reviews the prompt and provides a strength rating along with qualitative feedback to help the lab developer feel confident in their results.
Next Steps
Released in June 2025, this feature is already unlocking new possibilities for lab authors. I'll be tracking adoption and feedback through Pendo to guide future improvements.
