Figma · AI product · Mid · 45 to 60 minutes

Which AI job should improve finding design work?

Figma is a collaborative design and prototyping platform where teams create, share, and review interface and product design files. Answer this product management case and get it graded line by line against a weighted rubric.

Product strategy
User insight
AI product sense
Prioritization
Experiment design

The situation

A product leader responsible for how people discover work in Figma asks you to recommend the next AI investment for this area. Designers and nearby teams keep posting in Slack asking where the source file for a screenshot, mock, or component lives, which suggests people are losing real time hunting across files, teams, and inconsistent naming. At the same time, AI interest is high inside the company, and several ideas are competing for attention, including assistance while designing and intelligence that helps people retrieve the right file or content. Assume you can fund one cross functional team for two quarters. Your job is to decide which user problem to prioritize, whether AI is actually the right tool for it, and what would need to be true before launch. The tension is that the pain seems real, but the product shape is not obvious, and a weak result could damage trust because search and discovery in a work tool need a high quality bar.

What to produce

Write a 1 to 2 page product strategy memo.

Your answer must

State the primary user problem and target user, and explain why the observed behavior points to that problem rather than to a more general AI opportunity.
Compare at least two plausible AI use cases for this area, and recommend one to pursue now, or recommend waiting if you believe the case is not yet strong enough.
Describe an MVP, including the core user flow, what the system should and should not try to do at launch, and why.
Propose a validation plan for the next two quarters, including research, success metrics, quality thresholds, and explicit pivot or stop criteria.
Identify the main risks and dependencies, including trust, relevance quality, data or indexing needs, and access control or privacy concerns.

Constraints

You cannot expose any content a user does not already have permission to view.
You should not assume a large platform rewrite before proving user value.
Your recommendation must include a clear quality bar that would justify launch in a professional work tool.

Format

Use sections for problem, options, recommendation, MVP, validation, and risks.

How Figma actually played it

Figma explains how it went from exploring design autocomplete to launching visual and semantic search, after noticing designers losing time hunting for source files, shown by hundreds of Slack requests. The post covers the months of scoping and testing and the pivot that followed. It is a useful example of picking the AI use case from observed pain rather than from the technology.

Find the AI use case in observed behaviour, such as repeated help requests in Slack.
Be ready to pivot when an early concept does not fit the real job.
Obvious-in-hindsight features often take months of scoping.
Build for the infrastructure and quality bar the feature needs before launch.

Built from How we built AI search in Figma by Vincent van der Meulen, Figma.