Helping contributors understand the work before committing to it

The product split transcript discovery, claiming, editing and rewards across disconnected workflows, making it difficult for contributors to understand what to do next and leaving incomplete work locked for 24 hours.

I audited the end-to-end journey and reorganized it around clearer stages. Contributors could preview a transcript before claiming it, manage active and completed reviews separately, and withdraw work they could not complete.

The redesigned workflow was approved and handed off for development.

I walked through the live product as both a new and returning contributor, documenting friction across discovery, claiming, editing and review history.

I reviewed the findings with the team lead and developer to distinguish usability problems from intentional product rules and backend constraints.

Each appeared at a different point in the journey: finding work, committing to it, and tracking it. All three came from the same disconnect between screens.

The three problems shared one root cause: the product had the right pieces, but nothing connecting them. Instead of fixing each screen on its own, I reorganized the workflow into a sequence of stages-browsing, previewing, claiming, editing and tracking-that stayed connected as contributors moved through them.

The decisions below show how each stage was resolved.

The redesign had to preserve the existing claim and reward logic, use the established design system and minimize backend development. I prioritized changes to navigation, hierarchy and interaction while adapting or deferring ideas that required additional product logic.

Available Reviews and My Reviews represented the product's two primary contributor goals: finding new work and continuing existing work.

I moved both out of the profile drawer and into the primary navigation so contributors could access them throughout the review journey.

The original table separated related information across distant columns and repeated Claim on every row. I regrouped each transcript's title, source, speakers, labels, duration and reward into one selectable item that opened a preview.

This shifted the first action from accepting work to evaluating it.

The original reward indicator used both colour and coin quantity, making its levels harder to decode. I proposed displaying exact sats, but the team did not want payout to become the primary selection criterion.

I retained three relative reward levels while simplifying them into monochrome coin stacks.

Selecting a transcript now opens the editor in read-only preview mode. Contributors can inspect the source video, transcript, speakers, labels and review details before claiming it.

Claiming unlocks the editing controls and begins the existing 24-hour review window.

The editor now keeps claim status and its primary actions visible throughout the review. Contributors can withdraw work they cannot complete, allowing the transcript to return to the available pool instead of remaining unavailable until the claim expires.

I brought video navigation, timestamp controls and speaker management into the editing workspace. I also reduced competing toolbar actions, collapsed secondary transcript details and clarified the deadline treatment.

I proposed separating active work from review history because the two views supported different decisions. Active reviews prioritize deadlines and continuation actions, while history tracks pending, published, expired and rewarded work.

After reviewing alternative structures with the team, we moved forward with this direction.

The redesign was approved and handed off for development. Working inside a live product, most of my decisions were about restraint, not new ideas: I wanted the reward indicator to show exact sats, but the team didn't want payout to be the main reason someone picks a review, so I kept it relative. A lot of this project was telling apart what was broken from what was just a rule I had to work around.

The audit was based on my own walkthrough, not contributor interviews. Once this ships, I'd want to time how long reviews take in the new editor versus the old one, and check in with contributors on two things: whether the controls I brought into the editor actually made editing feel faster, and whether moving between stages felt smooth.