Initiative #255

Description of the initiative

We’re exploring the use of a SaaS design tool with built-in AI, such as Claude Design, to support UX and design work using unclassified information. The goal is to help with things like early ideation, drafting content, prototyping, and general design workflows to improve speed and consistency. We’re also trying to understand the leanest and most efficient way to go through AI governance for this type of tool, while still meeting privacy, security, and policy requirements.

Problem statement

Design work today involves a lot of manual and repetitive effort, like drafting content, creating variations, documenting decisions, which slows delivery and can lead to inconsistencies. There is also no clear or efficient way for evaluating and adopting low-risk AI tools, which creates uncertainty and can cause delays. This presents an opportunity to improve consistency and allow for more efficient design delivery within governance constraints.

Expected outcome if AI is selected

Success would mean reducing time spent on repetitive design tasks for content, documentation, etc. by ~20-30%, while enabling faster turnaround on design work. It would also improve consistency in outputs and allow teams to focus more on more design decisions.

Risks

Potential risks could be the accidental use of sensitive information in the tool, even if the intent is to use only unclassified data. There may also be uncertainty around how data is stored or retained by the SaaS provider. From a design perspective, there could be too much reliance on AI generated content, potentially impacting quality and accuracy if not reviewed carefully. There is also a risk of inconsistent use across teams without clear guidance or standards.

Strategic alignment

• Sustainable management of Canadian fisheries

Areas of impact

• Time

Feasibility

• ✓Data is available and usable